2011-06-30 17:46:53 +02:00
|
|
|
/* Copyright (c) 2008 MySQL AB, 2009 Sun Microsystems, Inc.
|
|
|
|
Use is subject to license terms.
|
2009-09-26 06:49:49 +02:00
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
the Free Software Foundation; version 2 of the License.
|
|
|
|
|
|
|
|
This program is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with this program; if not, write to the Free Software
|
|
|
|
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */
|
|
|
|
|
|
|
|
|
2017-06-18 05:42:16 +02:00
|
|
|
#include <my_global.h>
|
2009-09-26 06:49:49 +02:00
|
|
|
#include "semisync_slave.h"
|
MDEV-29369: rpl.rpl_semi_sync_shutdown_await_ack fails regularly with Result content mismatch
This test was prone to failures for a few reasons, summarized below:
1) MDEV-32168 introduced “only_running_threads=1” to
slave_stop.inc, which allowed the stop logic to bypass an
attempting-to-reconnect IO thread. That is, the IO thread could
realize the master shutdown in `read_event()`, and thereby call into
`try_to_reconnect()`. This would leave the IO thread up when the
test expected it to be stopped. Fixed by explicitly stopping the
IO thread and allowing an error state, as the above case would
lead to errno 2003.
2) On slow systems (or those running profiling tools, e.g. MSAN),
the waiting-for-ack transaction can complete before the system
processes the `SHUTDOWN WAIT FOR ALL SLAVES`. There was shutdown
preparation logic in-between the transaction and shutdown itself,
which contributes to this problem. This patch also moves this
preparation logic before the transaction, so there is less to do
in-between the calls.
3) Changed work-around for MDEV-28141 to use debug_sync instead
of sleep delay, as it was still possible to hit the bug on very
slow systems.
4) Masked MTR variable reset with disable/enable query log
Reviewed By:
============
Kristian Nielsen <knielsen@knielsen-hq.org>
2024-01-09 21:47:02 +01:00
|
|
|
#include "debug_sync.h"
|
2009-09-26 06:49:49 +02:00
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
Repl_semi_sync_slave repl_semisync_slave;
|
2017-11-22 16:10:34 +01:00
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
my_bool global_rpl_semi_sync_slave_enabled= 0;
|
2017-11-22 16:10:34 +01:00
|
|
|
char rpl_semi_sync_slave_delay_master;
|
2017-10-25 10:07:44 +02:00
|
|
|
ulong rpl_semi_sync_slave_trace_level;
|
2017-11-22 16:10:34 +01:00
|
|
|
unsigned int rpl_semi_sync_slave_kill_conn_timeout;
|
|
|
|
unsigned long long rpl_semi_sync_slave_send_ack = 0;
|
2009-09-26 06:49:49 +02:00
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
int Repl_semi_sync_slave::init_object()
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
|
|
|
int result= 0;
|
|
|
|
|
2017-11-27 20:06:17 +01:00
|
|
|
m_init_done = true;
|
2009-09-26 06:49:49 +02:00
|
|
|
|
|
|
|
/* References to the parameter works after set_options(). */
|
2017-11-25 17:54:42 +01:00
|
|
|
set_trace_level(rpl_semi_sync_slave_trace_level);
|
|
|
|
set_delay_master(rpl_semi_sync_slave_delay_master);
|
|
|
|
set_kill_conn_timeout(rpl_semi_sync_slave_kill_conn_timeout);
|
2009-09-26 06:49:49 +02:00
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
static bool local_semi_sync_enabled;
|
|
|
|
|
|
|
|
int rpl_semi_sync_enabled(THD *thd, SHOW_VAR *var, void *buff,
|
|
|
|
system_status_var *status_var,
|
|
|
|
enum_var_type scope)
|
|
|
|
{
|
|
|
|
local_semi_sync_enabled= repl_semisync_slave.get_slave_enabled();
|
|
|
|
var->type= SHOW_BOOL;
|
|
|
|
var->value= (char*) &local_semi_sync_enabled;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2021-01-26 01:20:05 +01:00
|
|
|
int Repl_semi_sync_slave::slave_read_sync_header(const uchar *header,
|
2017-11-25 17:54:42 +01:00
|
|
|
unsigned long total_len,
|
|
|
|
int *semi_flags,
|
2021-01-26 01:20:05 +01:00
|
|
|
const uchar **payload,
|
2017-11-25 17:54:42 +01:00
|
|
|
unsigned long *payload_len)
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
|
|
|
int read_res = 0;
|
2017-11-27 12:00:08 +01:00
|
|
|
DBUG_ENTER("Repl_semi_sync_slave::slave_read_sync_header");
|
2009-09-26 06:49:49 +02:00
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
if (get_slave_enabled())
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
2021-09-09 10:58:45 +02:00
|
|
|
if (!DBUG_IF("semislave_corrupt_log")
|
2021-01-26 01:20:05 +01:00
|
|
|
&& header[0] == k_packet_magic_num)
|
2017-11-22 16:10:34 +01:00
|
|
|
{
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
bool semi_sync_need_reply = (header[1] & k_packet_flag_sync);
|
2017-11-22 16:10:34 +01:00
|
|
|
*payload_len = total_len - 2;
|
|
|
|
*payload = header + 2;
|
|
|
|
|
2017-11-27 12:00:08 +01:00
|
|
|
DBUG_PRINT("semisync", ("%s: reply - %d",
|
|
|
|
"Repl_semi_sync_slave::slave_read_sync_header",
|
|
|
|
semi_sync_need_reply));
|
2017-11-22 16:10:34 +01:00
|
|
|
|
|
|
|
if (semi_sync_need_reply)
|
|
|
|
*semi_flags |= SEMI_SYNC_NEED_ACK;
|
2017-11-25 17:54:42 +01:00
|
|
|
if (is_delay_master())
|
2017-11-22 16:10:34 +01:00
|
|
|
*semi_flags |= SEMI_SYNC_SLAVE_DELAY_SYNC;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
sql_print_error("Missing magic number for semi-sync packet, packet "
|
|
|
|
"len: %lu", total_len);
|
|
|
|
read_res = -1;
|
|
|
|
}
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2017-11-22 16:10:34 +01:00
|
|
|
*payload= header;
|
|
|
|
*payload_len= total_len;
|
2009-09-26 06:49:49 +02:00
|
|
|
}
|
|
|
|
|
2017-11-27 12:00:08 +01:00
|
|
|
DBUG_RETURN(read_res);
|
2009-09-26 06:49:49 +02:00
|
|
|
}
|
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
/*
|
|
|
|
Set default semisync variables and print some replication info to the log
|
|
|
|
|
|
|
|
Note that the main setup is done in request_transmit()
|
|
|
|
*/
|
|
|
|
|
|
|
|
void Repl_semi_sync_slave::slave_start(Master_info *mi)
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
Set semi_sync_enabled at slave start. This is not changed until next
|
|
|
|
slave start or reconnect.
|
|
|
|
*/
|
|
|
|
bool semi_sync= global_rpl_semi_sync_slave_enabled;
|
|
|
|
|
|
|
|
set_slave_enabled(semi_sync);
|
|
|
|
mi->semi_sync_reply_enabled= 0;
|
2017-10-25 10:07:44 +02:00
|
|
|
|
2009-09-26 06:49:49 +02:00
|
|
|
sql_print_information("Slave I/O thread: Start %s replication to\
|
|
|
|
master '%s@%s:%d' in log '%s' at position %lu",
|
|
|
|
semi_sync ? "semi-sync" : "asynchronous",
|
2017-11-22 16:10:34 +01:00
|
|
|
const_cast<char *>(mi->user), mi->host, mi->port,
|
|
|
|
const_cast<char *>(mi->master_log_name),
|
|
|
|
(unsigned long)(mi->master_log_pos));
|
2009-09-26 06:49:49 +02:00
|
|
|
|
2017-11-22 16:10:34 +01:00
|
|
|
/*clear the counter*/
|
|
|
|
rpl_semi_sync_slave_send_ack= 0;
|
2009-09-26 06:49:49 +02:00
|
|
|
}
|
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
void Repl_semi_sync_slave::slave_stop(Master_info *mi)
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
2017-11-25 17:54:42 +01:00
|
|
|
if (get_slave_enabled())
|
MDEV-29369: rpl.rpl_semi_sync_shutdown_await_ack fails regularly with Result content mismatch
This test was prone to failures for a few reasons, summarized below:
1) MDEV-32168 introduced “only_running_threads=1” to
slave_stop.inc, which allowed the stop logic to bypass an
attempting-to-reconnect IO thread. That is, the IO thread could
realize the master shutdown in `read_event()`, and thereby call into
`try_to_reconnect()`. This would leave the IO thread up when the
test expected it to be stopped. Fixed by explicitly stopping the
IO thread and allowing an error state, as the above case would
lead to errno 2003.
2) On slow systems (or those running profiling tools, e.g. MSAN),
the waiting-for-ack transaction can complete before the system
processes the `SHUTDOWN WAIT FOR ALL SLAVES`. There was shutdown
preparation logic in-between the transaction and shutdown itself,
which contributes to this problem. This patch also moves this
preparation logic before the transaction, so there is less to do
in-between the calls.
3) Changed work-around for MDEV-28141 to use debug_sync instead
of sleep delay, as it was still possible to hit the bug on very
slow systems.
4) Masked MTR variable reset with disable/enable query log
Reviewed By:
============
Kristian Nielsen <knielsen@knielsen-hq.org>
2024-01-09 21:47:02 +01:00
|
|
|
{
|
|
|
|
#ifdef ENABLED_DEBUG_SYNC
|
|
|
|
/*
|
|
|
|
TODO: Remove after MDEV-28141
|
|
|
|
*/
|
|
|
|
DBUG_EXECUTE_IF("delay_semisync_kill_connection_for_mdev_28141", {
|
|
|
|
const char act[]= "now "
|
|
|
|
"signal at_semisync_kill_connection "
|
|
|
|
"wait_for continue_semisync_kill_connection";
|
|
|
|
DBUG_ASSERT(debug_sync_service);
|
|
|
|
DBUG_ASSERT(!debug_sync_set_action(mi->io_thd, STRING_WITH_LEN(act)));
|
|
|
|
};);
|
|
|
|
#endif
|
2017-11-25 17:54:42 +01:00
|
|
|
kill_connection(mi->mysql);
|
MDEV-29369: rpl.rpl_semi_sync_shutdown_await_ack fails regularly with Result content mismatch
This test was prone to failures for a few reasons, summarized below:
1) MDEV-32168 introduced “only_running_threads=1” to
slave_stop.inc, which allowed the stop logic to bypass an
attempting-to-reconnect IO thread. That is, the IO thread could
realize the master shutdown in `read_event()`, and thereby call into
`try_to_reconnect()`. This would leave the IO thread up when the
test expected it to be stopped. Fixed by explicitly stopping the
IO thread and allowing an error state, as the above case would
lead to errno 2003.
2) On slow systems (or those running profiling tools, e.g. MSAN),
the waiting-for-ack transaction can complete before the system
processes the `SHUTDOWN WAIT FOR ALL SLAVES`. There was shutdown
preparation logic in-between the transaction and shutdown itself,
which contributes to this problem. This patch also moves this
preparation logic before the transaction, so there is less to do
in-between the calls.
3) Changed work-around for MDEV-28141 to use debug_sync instead
of sleep delay, as it was still possible to hit the bug on very
slow systems.
4) Masked MTR variable reset with disable/enable query log
Reviewed By:
============
Kristian Nielsen <knielsen@knielsen-hq.org>
2024-01-09 21:47:02 +01:00
|
|
|
}
|
2021-10-21 04:13:45 +02:00
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
set_slave_enabled(0);
|
2009-09-26 06:49:49 +02:00
|
|
|
}
|
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
void Repl_semi_sync_slave::slave_reconnect(Master_info *mi)
|
2009-09-26 06:49:49 +02:00
|
|
|
{
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
/*
|
|
|
|
Start semi-sync either if it globally enabled or if was enabled
|
|
|
|
before the reconnect.
|
|
|
|
*/
|
|
|
|
if (global_rpl_semi_sync_slave_enabled || get_slave_enabled())
|
|
|
|
slave_start(mi);
|
2009-09-26 06:49:49 +02:00
|
|
|
}
|
2017-10-25 10:07:44 +02:00
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
void Repl_semi_sync_slave::kill_connection(MYSQL *mysql)
|
2017-11-22 16:10:34 +01:00
|
|
|
{
|
|
|
|
if (!mysql)
|
|
|
|
return;
|
2017-10-25 10:07:44 +02:00
|
|
|
|
2017-11-22 16:10:34 +01:00
|
|
|
char kill_buffer[30];
|
|
|
|
MYSQL *kill_mysql = NULL;
|
2021-10-21 04:13:45 +02:00
|
|
|
size_t kill_buffer_length;
|
|
|
|
|
2017-11-22 16:10:34 +01:00
|
|
|
kill_mysql = mysql_init(kill_mysql);
|
2017-11-27 20:06:17 +01:00
|
|
|
mysql_options(kill_mysql, MYSQL_OPT_CONNECT_TIMEOUT, &m_kill_conn_timeout);
|
|
|
|
mysql_options(kill_mysql, MYSQL_OPT_READ_TIMEOUT, &m_kill_conn_timeout);
|
|
|
|
mysql_options(kill_mysql, MYSQL_OPT_WRITE_TIMEOUT, &m_kill_conn_timeout);
|
2017-10-25 10:07:44 +02:00
|
|
|
|
2017-11-22 16:10:34 +01:00
|
|
|
bool ret= (!mysql_real_connect(kill_mysql, mysql->host,
|
|
|
|
mysql->user, mysql->passwd,0, mysql->port, mysql->unix_socket, 0));
|
2021-09-09 10:58:45 +02:00
|
|
|
if (DBUG_IF("semisync_slave_failed_kill") || ret)
|
2017-11-22 16:10:34 +01:00
|
|
|
{
|
|
|
|
sql_print_information("cannot connect to master to kill slave io_thread's "
|
|
|
|
"connection");
|
2021-10-21 04:13:45 +02:00
|
|
|
goto failed_graceful_kill;
|
|
|
|
}
|
|
|
|
|
|
|
|
kill_buffer_length= my_snprintf(kill_buffer, 30, "KILL %lu",
|
|
|
|
mysql->thread_id);
|
|
|
|
if (mysql_real_query(kill_mysql, kill_buffer, (ulong)kill_buffer_length))
|
|
|
|
{
|
|
|
|
sql_print_information(
|
|
|
|
"Failed to gracefully kill our active semi-sync connection with "
|
|
|
|
"primary. Silently closing the connection.");
|
|
|
|
goto failed_graceful_kill;
|
2017-11-22 16:10:34 +01:00
|
|
|
}
|
2021-10-21 04:13:45 +02:00
|
|
|
|
|
|
|
end:
|
2017-11-22 16:10:34 +01:00
|
|
|
mysql_close(kill_mysql);
|
2021-10-21 04:13:45 +02:00
|
|
|
return;
|
|
|
|
|
|
|
|
failed_graceful_kill:
|
|
|
|
/*
|
|
|
|
If we fail to issue `KILL` on the primary to kill the active semi-sync
|
|
|
|
connection; we need to locally clean up our side of the connection. This
|
|
|
|
is because mysql_close will send COM_QUIT on the active semi-sync
|
|
|
|
connection, causing the primary to error.
|
|
|
|
*/
|
|
|
|
net_clear(&(mysql->net), 0);
|
|
|
|
end_server(mysql);
|
|
|
|
goto end;
|
2017-10-25 10:07:44 +02:00
|
|
|
}
|
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
int Repl_semi_sync_slave::request_transmit(Master_info *mi)
|
2017-10-25 10:07:44 +02:00
|
|
|
{
|
2017-11-22 16:10:34 +01:00
|
|
|
MYSQL *mysql= mi->mysql;
|
2017-10-25 10:07:44 +02:00
|
|
|
MYSQL_RES *res= 0;
|
|
|
|
MYSQL_ROW row;
|
|
|
|
const char *query;
|
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
if (!get_slave_enabled())
|
2017-10-25 10:07:44 +02:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
query= "SHOW VARIABLES LIKE 'rpl_semi_sync_master_enabled'";
|
2018-02-06 13:55:58 +01:00
|
|
|
if (mysql_real_query(mysql, query, (ulong)strlen(query)) ||
|
2017-10-25 10:07:44 +02:00
|
|
|
!(res= mysql_store_result(mysql)))
|
|
|
|
{
|
2017-11-22 16:10:34 +01:00
|
|
|
sql_print_error("Execution failed on master: %s, error :%s", query, mysql_error(mysql));
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
set_slave_enabled(0);
|
2017-10-25 10:07:44 +02:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
row= mysql_fetch_row(res);
|
2024-02-01 18:36:14 +01:00
|
|
|
if (DBUG_IF("master_not_support_semisync") || (!row || ! row[1]))
|
2017-10-25 10:07:44 +02:00
|
|
|
{
|
|
|
|
/* Master does not support semi-sync */
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
if (!row)
|
|
|
|
sql_print_warning("Master server does not support semi-sync, "
|
|
|
|
"fallback to asynchronous replication");
|
|
|
|
set_slave_enabled(0);
|
2017-10-25 10:07:44 +02:00
|
|
|
mysql_free_result(res);
|
|
|
|
return 0;
|
|
|
|
}
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
if (strcmp(row[1], "ON"))
|
|
|
|
sql_print_information("Slave has semi-sync enabled but master server does "
|
|
|
|
"not. Semi-sync will be activated when master "
|
|
|
|
"enables it");
|
2017-10-25 10:07:44 +02:00
|
|
|
mysql_free_result(res);
|
|
|
|
|
|
|
|
/*
|
2017-11-22 16:10:34 +01:00
|
|
|
Tell master dump thread that we want to do semi-sync
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
replication. This is done by setting a thread local variable in
|
|
|
|
the master connection.
|
2017-10-25 10:07:44 +02:00
|
|
|
*/
|
|
|
|
query= "SET @rpl_semi_sync_slave= 1";
|
2018-02-06 13:55:58 +01:00
|
|
|
if (mysql_real_query(mysql, query, (ulong)strlen(query)))
|
2017-10-25 10:07:44 +02:00
|
|
|
{
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
sql_print_error("%s on master failed", query);
|
|
|
|
set_slave_enabled(0);
|
2017-10-25 10:07:44 +02:00
|
|
|
return 1;
|
|
|
|
}
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
mi->semi_sync_reply_enabled= 1;
|
|
|
|
/* Inform net_server that pkt_nr can come out of order */
|
|
|
|
mi->mysql->net.pkt_nr_can_be_reset= 1;
|
2017-10-25 10:07:44 +02:00
|
|
|
mysql_free_result(mysql_store_result(mysql));
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-11-25 17:54:42 +01:00
|
|
|
int Repl_semi_sync_slave::slave_reply(Master_info *mi)
|
2017-10-25 10:07:44 +02:00
|
|
|
{
|
2017-11-22 16:10:34 +01:00
|
|
|
MYSQL* mysql= mi->mysql;
|
|
|
|
const char *binlog_filename= const_cast<char *>(mi->master_log_name);
|
|
|
|
my_off_t binlog_filepos= mi->master_log_pos;
|
|
|
|
NET *net= &mysql->net;
|
|
|
|
uchar reply_buffer[REPLY_MAGIC_NUM_LEN
|
|
|
|
+ REPLY_BINLOG_POS_LEN
|
|
|
|
+ REPLY_BINLOG_NAME_LEN];
|
|
|
|
int reply_res = 0;
|
2018-02-06 13:55:58 +01:00
|
|
|
size_t name_len = strlen(binlog_filename);
|
2017-11-27 12:00:08 +01:00
|
|
|
DBUG_ENTER("Repl_semi_sync_slave::slave_reply");
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
DBUG_ASSERT(get_slave_enabled() && mi->semi_sync_reply_enabled);
|
2017-10-25 10:07:44 +02:00
|
|
|
|
MDEV-32551: "Read semi-sync reply magic number error" warnings on master
rpl_semi_sync_slave_enabled_consistent.test and the first part of
the commit message comes from Brandon Nesterenko.
A test to show how to induce the "Read semi-sync reply magic number
error" message on a primary. In short, if semi-sync is turned on
during the hand-shake process between a primary and replica, but
later a user negates the rpl_semi_sync_slave_enabled variable while
the replica's IO thread is running; if the io thread exits, the
replica can skip a necessary call to kill_connection() in
repl_semisync_slave.slave_stop() due to its reliance on a global
variable. Then, the replica will send a COM_QUIT packet to the
primary on an active semi-sync connection, causing the magic number
error.
The test in this patch exits the IO thread by forcing an error;
though note a call to STOP SLAVE could also do this, but it ends up
needing more synchronization. That is, the STOP SLAVE command also
tries to kill the VIO of the replica, which makes a race with the IO
thread to try and send the COM_QUIT before this happens (which would
need more debug_sync to get around). See THD::awake_no_mutex for
details as to the killing of the replica’s vio.
Notes:
- The MariaDB documentation does not make it clear that when one
enables semi-sync replication it does not matter if one enables
it first in the master or slave. Any order works.
Changes done:
- The rpl_semi_sync_slave_enabled variable is now a default value for
when semisync is started. The variable does not anymore affect
semisync if it is already running. This fixes the original reported
bug. Internally we now use repl_semisync_slave.get_slave_enabled()
instead of rpl_semi_sync_slave_enabled. To check if semisync is
active on should check the @@rpl_semi_sync_slave_status variable (as
before).
- The semisync protocol conflicts in the way that the original
MySQL/MariaDB client-server protocol was designed (client-server
send and reply packets are strictly ordered and includes a packet
number to allow one to check if a packet is lost). When using
semi-sync the master and slave can send packets at 'any time', so
packet numbering does not work. The 'solution' has been that each
communication starts with packet number 1, but in some cases there
is still a chance that the packet number check can fail. Fixed by
adding a flag (pkt_nr_can_be_reset) in the NET struct that one can
use to signal that packet number checking should not be done. This
is flag is set when semi-sync is used.
- Added Master_info::semi_sync_reply_enabled to allow one to configure
some slaves with semisync and other other slaves without semisync.
Removed global variable semi_sync_need_reply that would not work
with multi-master.
- Repl_semi_sync_master::report_reply_packet() can now recognize
the COM_QUIT packet from semisync slave and not give a
"Read semi-sync reply magic number error" error for this case.
The slave will be removed from the Ack listener.
- On Windows, don't stop semisync Ack listener just because one
slave connection is using socket_id > FD_SETSIZE.
- Removed busy loop in Ack_receiver::run() by using
"Self-pipe trick" to signal new slave and stop Ack_receiver.
- Changed some Repl_semi_sync_slave functions that always returns 0
from int to void.
- Added Repl_semi_sync_slave::slave_reconnect().
- Removed dummy_function Repl_semi_sync_slave::reset_slave().
- Removed some duplicate semisync notes from the error log.
- Add test of "if (get_slave_enabled() && semi_sync_need_reply)"
before calling Repl_semi_sync_slave::slave_reply().
(Speeds up the code as we can skip all initializations).
- If epl_semisync_slave.slave_reply() fails, we disable semisync
for that connection.
- We do not call semisync.switch_off() if there are no active slaves.
Instead we check in Repl_semi_sync_master::commit_trx() if there are
no active threads. This simplices the code.
- Changed assert() to DBUG_ASSERT() to ensure that the DBUG log is
flushed in case of asserts.
- Removed the internal rpl_semi_sync_slave_status as it is not needed
anymore. The @@rpl_semi_sync_slave_status status variable is now
mapped to rpl_semi_sync_enabled.
- Removed rpl_semi_sync_slave_enabled as it is not needed anymore.
Repl_semi_sync_slave::get_slave_enabled() contains the active status.
- Added checking that we do not add a slave twice with
Ack_receiver::add_slave(). This could happen with old code.
- Removed Repl_semi_sync_master::check_and_switch() as it is not
needed anymore.
- Ensure that when we call Ack_receiver::remove_slave() that the slave
is removed from the listener before function returns.
- Call listener.listen_on_sockets() outside of mutex for better
performance and less contested mutex.
- Ensure that listening is ignoring newly added slaves when checking for
responses.
- Fixed the master ack_receiver listener is not killed if there are no
connected slaves (and thus stop semisync handling of future
connections). This could happen if all slaves sockets where would be
marked as unreliable.
- Added unlink() to base_ilist_iterator and remove() to
I_List_iterator. This enables us to remove 'dead' slaves in
Ack_recever::run().
- kill_zombie_dump_threads() now does killing of dump threads properly.
- It can now kill several threads (should be impossible but could
happen if IO slaves reconnects very fast).
- We now wait until the dump thread is done before starting the
dump.
- Added an error if kill_zombie_dump_threads() fails.
- Set thd->variables.server_id before calling
kill_zombie_dump_threads(). This simplies the code.
- Added a lot of comments both in code and tests.
- Removed DBUG_EVALUATE_IF "failed_slave_start" as it is not used.
Test changes:
- rpl.rpl_session_var2 added which runs rpl.rpl_session_var test with
semisync enabled.
- Some timings changed slight with startup of slave which caused
rpl_binlog_dump_slave_gtid_state_info.text to fail as it checked the
error log file before the slave had started properly. Fixed by
adding wait_for_pattern_in_file.inc that allows waiting for the
pattern to appear in the log file.
- Tests have been updated so that we first set
rpl_semi_sync_master_enabled on the master and then set
rpl_semi_sync_slave_enabled on the slaves (this is according to how
the MariaDB documentation document how to setup semi-sync).
- Error text "Master server does not have semi-sync enabled" has been
replaced with "Master server does not support semi-sync" for the
case when the master supports semi-sync but semi-sync is not
enabled.
Other things:
- Some trivial cleanups in Repl_semi_sync_master::update_sync_header().
- We should in 11.3 changed the default value for
rpl-semi-sync-master-wait-no-slave from TRUE to FALSE as the TRUE
does not make much sense as default. The main difference with using
FALSE is that we do not wait for semisync Ack if there are no slave
threads. In the case of TRUE we wait once, which did not bring any
notable benefits except slower startup of master configured for
using semisync.
Co-author: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This solves the problem reported in MDEV-32960 where a new
slave may not be registered in time and the master disables
semi sync because of that.
2023-11-09 00:57:58 +01:00
|
|
|
/* Prepare the buffer of the reply. */
|
|
|
|
reply_buffer[REPLY_MAGIC_NUM_OFFSET] = k_packet_magic_num;
|
|
|
|
int8store(reply_buffer + REPLY_BINLOG_POS_OFFSET, binlog_filepos);
|
|
|
|
memcpy(reply_buffer + REPLY_BINLOG_NAME_OFFSET,
|
|
|
|
binlog_filename,
|
|
|
|
name_len + 1 /* including trailing '\0' */);
|
|
|
|
|
|
|
|
DBUG_PRINT("semisync", ("%s: reply (%s, %lu)",
|
|
|
|
"Repl_semi_sync_slave::slave_reply",
|
|
|
|
binlog_filename, (ulong)binlog_filepos));
|
|
|
|
|
|
|
|
/*
|
|
|
|
We have to do a net_clear() as with semi-sync the slave_reply's are
|
|
|
|
interleaved with data from the master and then the net->pkt_nr
|
|
|
|
cannot be kept in sync. Better to start pkt_nr from 0 again.
|
|
|
|
*/
|
|
|
|
net_clear(net, 0);
|
|
|
|
/* Send the reply. */
|
|
|
|
reply_res = my_net_write(net, reply_buffer,
|
|
|
|
name_len + REPLY_BINLOG_NAME_OFFSET);
|
|
|
|
if (!reply_res)
|
2017-11-22 16:10:34 +01:00
|
|
|
{
|
2024-02-01 18:36:14 +01:00
|
|
|
reply_res= DBUG_IF("semislave_failed_net_flush") || net_flush(net);
|
2017-11-22 16:10:34 +01:00
|
|
|
if (!reply_res)
|
|
|
|
rpl_semi_sync_slave_send_ack++;
|
|
|
|
}
|
2017-11-27 12:00:08 +01:00
|
|
|
DBUG_RETURN(reply_res);
|
2017-10-25 10:07:44 +02:00
|
|
|
}
|