If the slave gets a reconnect in the middle of a GTID event group, normally
it will re-fetch that event group, skipping the first part that was already
queued for the SQL thread.
However, if the master crashed while writing the event group, the group is
incomplete. This patch detects this case and makes sure that the
transaction is rolled back and nothing is skipped from any following
event groups.
Similarly, a network proxy might cause the reconnect to end up on a
different master server. Detect this by noticing a different server_id,
and similarly in this case roll back the partially received group.
Item_string::eq() and Item_param::eq() in string context behaved differently.
Introducing a new class Item_basic_value to share the eq() code between
literals (Item_int, Item_double, Item_string, Item_null) and Item_param.
MDEV-6666 Malformed result for CONCAT(utf8_column, binary_string)
Item_static_string_func::safe_charset_converter() and
Item_hex_string::safe_charset_converter() did not
handle character sets with mbminlen>1 properly, as well as
did not handle conversion from binary to multi-byte well.
Introducing Item::const_charset_converter(), to reuse it in a number
of Item_*::safe_charset_converter().
Analysis: When database is migrated from 5.5 or earlier and
database needs crash recovery, there is possibility that
SYS_DATAFILES system table does not exists, but
crash recovery in function dict_check_tablespaces_and_store_max_id()
assumes that SYS_DATAFILES exists.
Fix: If SYS_DATAFILES does not exists, create it before
we end up to function dict_check_tablespaces_and_store_max_id()
on crash recovery.
After-review fixes.
Mainly catching if the wait in wait_for_workers_idle() is aborted due to kill.
In this case, we should return an error and not proceed to execute the format
description event, as other threads might still be running for a bit until the
error is caught in all threads.
Follow-up patch, fixing a possible deadlock issue.
If the master crashes in the middle of an event group, there can be an active
transaction in a worker thread when we encounter the following master restart
format description event. In this case, we need to notify that worker thread
to abort and roll back the partial event group. Otherwise a deadlock occurs:
the worker thread waits for the commit that never arrives, and the SQL driver
thread waits for the worker thread to complete its event group, which it never
does.
When a master server starts up, it logs a special format_description event at
the start of a new binlog to mark that is has restarted. This is used by a
slave to drop all temporary tables - this is needed in case the master crashed
and did not have a chance to send explicit DROP TEMPORARY TABLE statements to
the slave.
In parallel replication, we need to be careful when dropping the temporary
tables - we need to be sure that no prior events are still executing that
might be using the temporary tables to be dropped, _and_ that no following
events have started executing that might have created new temporary tables
that should not be dropped.
This was not handled correctly, which could cause errors about access to not
existing temporary tables or even crashes. This patch implements that such
format_description events cause serialisation of event execution; all prior
events are executed to completion first, then the format_description event is
executed, dropping temporary tables, then following events are queued for
execution.
Master restarts should be sufficiently infrequent that the resulting loss of
parallelism should be of minimal impact.
Part of this work is based on Stewart Smitch's memory barrier and lower priori
patches for power8.
- Added memory syncronization for innodb & xtradb for power8.
- Added HAVE_WINDOWS_MM_FENCE to CMakeList.txt
- Added os_isync to fix a syncronization problem on power
- Added log_get_lsn_nowait which is now used srv_error_monitor_thread to ensur
if log mutex is locked.
All changes done both for InnoDB and Xtradb
The problem occured when using parallel replication, and an error occured that
caused the SQL thread to stop when the IO thread had already reached a
following binlog file from the master (or otherwise performed a relay log
rotation).
In this case, the Rotate Event at the end of the relay log file could still be
executed, even though an earlier event in that relay log file had gotten an
error. This would cause the position to be incorrectly updated, so that upon
restart of the SQL thread, the event that had failed would be silently skipped
and ignored, causing replication corruption.
Fixed by checking before executing Rotate Event, whether an earlier event
has failed. If so, the Rotate Event is not executed, just dequeued, same as
for other normal events following a failing event.
If one uses 4 --verbose to mysql_upgrade it will also write out all mysqlcheck commands invoked.
mysql-test/r/mysql_upgrade.result:
Updated results from changing phases
mysql-test/r/mysql_upgrade_no_innodb.result:
dated results from changing phases
mysql-test/r/mysql_upgrade_ssl.result:
dated results from changing phases
The bug was that in some cases, if a replicated transaction was rolled back
due to deadlock, during the subsequent retry of that transaction, the
gtid_slave_pos would _not_ be updated with the new GTID, leaving the GTID
position of the slave incorrect.
Fix this by ensuring during the retry that we clear the flag that marks that
the GTID has already been recorded in gtid_slave_pos, so that the update of
gtid_slave_pos will be done again during the retry.
In the original bug, the symptom was an assertion due to OPTION_GTID_BEGIN not
being cleared during the retry of the transaction. The reason was some code in
handling of a COMMIT query event, which would not clear the flag when not
recording a GTID in gtid_slave_pos. This commit also fixes that code to always
clear the OPTION_GTID_BEGIN flag for clarity, though it is actually not
possible for OPTION_GTID_BEGIN to become set unless a GTID is pending for
update (after fixing the bug described above).
sporadically
Fix: Modify test to be smaller so that testcase timeout does not
trigger. We already have a test for --big-test setup
(innodb.innodb_simulate_comp_failures).
sporadically
Fix: Modify test to be smaller so that testcase timeout does not
trigger. We already have a test for --big-test setup
(innodb.innodb_simulate_comp_failures).
config.h.cmake:
define NOMINMAX, otherwise Windows system headers define min() and max() macros
sql/slave.cc:
mi->report() has one more argument in MariaDB
storage/xtradb/buf/buf0flu.cc:
xtradb fixes for windows, again
When a view is merged, mark its select_lex as already optimized, otherwise
its where clause (which doesn't have to be a valid Item after merging)
might be accessed later. But don't do that for inserts (where a view cannot
be simply merged, if one later needs to insert into it).
have been created by the last addition of new CONNECT features.
The version previous to this one is a preliminary test version and
should not be distributed.
- Handle indexed UPDATE/DELETE. Previously this was just tested and
an error message send when it could not be done. Now CONNECT can
do it in all the cases. It is done by a MRR like tchnique by making
a list of all update or delete to do, sort them, then execute them.
modified:
storage/connect/array.cpp
storage/connect/array.h
storage/connect/filamap.cpp
storage/connect/filamap.h
storage/connect/filamdbf.cpp
storage/connect/filamfix.cpp
storage/connect/filamfix.h
storage/connect/filamtxt.cpp
storage/connect/filamtxt.h
storage/connect/filamvct.cpp
storage/connect/filamvct.h
storage/connect/filamzip.cpp
storage/connect/filamzip.h
storage/connect/global.h
storage/connect/ha_connect.cc
storage/connect/ha_connect.h
- Differenciate Cardinality that returns a true or estimated table size
and GetMaxSize that return a value equal or greater than the table
row number. This fixes the errors of non matching opt files.
modified:
storage/connect/connect.cc
storage/connect/tabdos.cpp
storage/connect/tabdos.h
storage/connect/tabfix.cpp
storage/connect/table.cpp
storage/connect/tabmac.h
storage/connect/tabmysql.cpp
storage/connect/tabmysql.h
storage/connect/tabodbc.cpp
storage/connect/tabodbc.h
storage/connect/tabpivot.h
storage/connect/tabtbl.cpp
storage/connect/tabtbl.h
storage/connect/tabutil.cpp
storage/connect/tabutil.h
storage/connect/tabwmi.h
storage/connect/xtable.h
- Fix some errors and issues when making index and opt files.
Erase opt and index files for void tables.
Fix wrong calculation of Block and Last in MakeBlockValues.
Invalidate indexes before making opt file.
Fully handle blocked variable tables. Make opt file for blocked
variable tables even when they have no optimised colums.
modified:
storage/connect/tabdos.cpp
storage/connect/xindex.h
- Fix some errors making index
Return an error when the allocation is too small (should not
really occur now that GetMaxSize is sure)
Don't use XXROW index for DBF tables because of soft deleted lines.
modified:
storage/connect/xindex.cpp
- Typo
modified:
storage/connect/macutil.cpp
storage/connect/tabdos.h
storage/connect/tabsys.cpp
storage/connect/tabsys.h
A new command line option "galera-sst-mode" was introduced
in mysqldump as part of fix for MDEV-6316. But, since the
fix was pushed to maria-10.0-galera branch, the mysqldump
tool supplied with mariadb client deb/rpm packages, does not
have this new opion.
This fix contains the same patch along with a test case.
using "MariaDB-Galera-server"
rpm: Updated components' requires-list to include "MariaDB"
instead of "MariaDB-server".
deb: Added (ORed) 'mariadb-galera-server-10.0' to the 'Depends'
list of connect and oqgraph engine packages.
INSTALL SONAME ignores attempts to load the same plugin twice, it's not an error
(because one can load one plugin by name and then install soname for the rest).
But Federated and FederatedX are different plugins, despite having the same name.
Now plugin_add() only considers two plugins identical if their names are the same
string (compared as pointers). Otherwise it reports an error.,
