- innodb.doublewrite_debug should avoid the checkpoint
before killing the server. So used debug sync and
innodb_flush_sync to avoid the checkpoint completely.
Test case allowed to skip on MSAN builder due to extra
checkpoint.
Attempting to set a SAVEPOINT when one of the involved storage engines
does not support savepoints, raises an error, and results in statement
rollback. If Galera is enabled with binlog emulation, the above
scenario was not handled correctly, and resulted in cluster wide
inconsistency.
The problem was in wsrep_register_binlog_handler(), which is called
towards the beginning of SAVEPOINT execution. This function is
supposed to mark the beginning of statement position in trx cache
through `set_prev_position()`. However, it did so only on condition
that `get_prev_position()` returns `MY_OFF_T_UNDEF`.
This before statement position is typically reset to undefined at the
end of statement in `binlog_commit()` / `binlog_rollback()`.
However that's not the case with Galera and binlog emulation, for
which binlog commit / rollback hooks are not called due to the
optimization that avoids internal 2PC (MDEV-16509).
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
This commit fixes GTID inconsistency which was injected by mariabackup SST.
Donor node now writes new info file: donor_galera_info, which is streamed
along the mariabackup donation to the joiner node. The donor_galera_info
file contains both GTID and gtid domain_id, and joiner will use these to
initialize the GTID state.
Commit has new mtr test case: galera_3nodes.galera_gtid_consistency, which
exercises potentially harmful mariabackup SST scenarios. The test has also
scenario with IST joining.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
In spider_db_mbase_util::print_item_func(), if the sql item_func has
an UNKNOWN_FUNC type, by default the spider group by handler (gbh)
transform infix to prefix. But regexp should remain infix, so we add
an if condition to account for this.
This patch fixes cases where a transaction caused empty writeset to be
replicated. This could happen in the case where a transaction executes
a statement that initially manages to modify some data and therefore
appended keys some for certification. The statement is however rolled
back at some later stage due to some error (for example, a duplicate
key error). After statement rollback the transaction is still alive,
has no other changes. When committing such transaction, an empty
writeset was replicated through Galera.
The fix is to avoid calling into commit hook only when transaction
has appended one or keys for certification *and* has some data in
binlog cache to replicate. Otherwise, the commit is considered empty,
and goes through usual empty commit path.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Remove DB_LOCK_WAIT return code check as it should have been resolved to
one of the other errors by that point.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
The previous patch for MDEV-10653 changes the rpl_parallel::workers_idle()
function to use Relay_log_info::last_inuse_relaylog to check for idle
workers. But the code was missing a NULL check. Also, there was one place
during SQL slave thread start which was missing mutex synchronisation when
updating inuse_relaylog.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
The error-injection inject_mdev8031 simulates a deadlock kill in a specific
place, by setting killed_for_retry to RETRY_KILL_KILLED directly. If a real
deadlock kill triggers at the same time, it is possible for the thread to
complete its transaction retry and set rgi_slave to NULL before the real
readlock kill can complete in the background. This will cause a segfault
due to null-pointer access.
Fix by changing the error injection to do a real background deadlock kill,
which ensures that the thread will wait for any pending background kills to
complete.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
Item::val_str() sets the Item::null_value flag, so call it before checking
the flag, not after.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
trx_purge_truncate_history(): Release buf_pool.flush_list_mutex
before starting a rescan of buf_pool.flush_list, to ensure that
the buf_flush_page_cleaner thread (which may be holding buf_pool.mutex)
will be able to proceed.
This fixes up commit a0f02f7438 (MDEV-32757).
Tested by: Axel Schwenke
Reviewed by: Vladislav Lesin
opt_kill_long_query_type being an enum could be 0 corresponding
to ALL. When ALL is specified, the CONNECTION ADMIN is still
required.
Also check REPLICA MONITOR privilege and make the tests
find the results by recording stderr.
Noticed thanks to bug report by Tim van Dijen.
Fixes: 79b58f1ca8
Consider this query
SELECT t1.* FROM t1, (SELECT t2.b FROM t2 WHERE NOT EXISTS
(SELECT 1 FROM t3) GROUP BY b) sq where sq.b = t1.a;
If SELECT 1 FROM t3 is expensive, for example t3 has >
thd->variables.expensive_subquery_limit, first evaluation is deferred to
mysql_derived_fill(). There it is noted that, in the above case
NOT EXISTS (SELECT 1 FROM t3) is constant and false.
This causes the join variable zero_result_cause to be set to
"Impossible WHERE noticed after reading const tables" and the handler
for this join is never "opened" via handler::ha_open.
When mysql_derived_fill() is called for the next group of results, this
unopened handler is not taken into account.
reviewed by Igor Babaev (igor@mariadb.com)
row_ins_clust_index_entry_low(): Invoke btr_set_instant() in the same
mini-transaction that has successfully inserted the metadata record.
In this way, if inserting the metadata record fails before any
undo log record was written for it, the index root page will remain
consistent.
innobase_instant_try(): Remove the btr_set_instant() call.
Reviewed by: Thirunarayanan Balathandayuthapani
Tested by: Matthias Leich
Like all IF NOT EXISTS syntax, a Note should be generated.
The original commit of Seqeuences cleared the IF NOT EXISTS part
in the sql/sql_yacc.yy with lex->create_info.init(). Without this
bit set there was no way it could do anything other than error.
To remedy this removal, the sql_yacc.yy components have been
minimised as they where all set at the beginning of the ALTER.
This way the opt_if_not_exists correctly set the IF_EXISTS flag.
In MDEV-13005 (bb4dd70e7c) the error code changed, requiring
ER_UNKNOWN_SEQUENCES to be handled in the function
No_such_table_error_handler::handle_condition.
Allow for a CI system to be almost out of space, or having so
little use, that the Total space is the same as available or used.
Thanks Otto Kekäläinen for the bug report and testing.
This is a port of the Percona Server commit 5265f42e290573e9591f8ca28ab66afc051f89a3
which is the same as their bug PXB-1807: xtrabackup does not accept fractional values for
innodb_max_dirty_pages_pct
Problem:
Variable specified as double in MySQL server, but read as long in the
xtrabackup. This causes xtrabackup to fail at startup when the value
contains decimal point.
Fix:
Make xtrabackup to interpret the value as double to be compatible with
server.
calculate auto-inc value even if long duplicate check fails -
this is what the engine does for normal uniques.
auto-inc value is needed if it's a REPLACE
ha_innobase::check_if_supported_inplace_alter(): On ALTER_OPTIONS,
if innodb_file_per_table=1 and the table resides in the system tablespace,
require that the table be rebuilt (and moved to an .ibd file).
Reviewed by: Thirunarayanan Balathandayuthapani
Tested by: Matthias Leich
This test was using a sleep of 1 second in an attempt to ensure that the
timestamp that is part of an InnoDB status string would increase.
This not only prolongs the test execution time by 1+1 seconds, but it
also is inaccurate. It is possible that the actual sleep duration is
less than a second.
Let us wait for the creation of the file ib_buffer_pool and then wait
for the buffer pool dump completion. In that way, the test can complete
in a dozen or two milliseconds (1% of the previous duration) and work
more reliably.
Add OPTION_GTID_BEGIN to applying side thread. This is needed to avoid
intermediate commits when CREATE TABLE AS SELECT is applied, causing
one more GTID to be consumed with respect to executing node.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Return an error if user attempts to use SEQUENCEs in combination with
streaming replication in a Galera cluster. This is currently not
supported.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
AKA rpl.rpl_parallel, binlog_encryption.rpl_parallel fails in
buildbot with timeout in include
A replication parallel worker thread can deadlock with another
connection running SHOW SLAVE STATUS. That is, if the replication
worker thread is in do_gco_wait() and is killed, it will already
hold the LOCK_parallel_entry, and during error reporting, try to
grab the err_lock. SHOW SLAVE STATUS, however, grabs these locks in
reverse order. It will initially grab the err_lock, and then try to
grab LOCK_parallel_entry. This leads to a deadlock when both threads
have grabbed their first lock without the second.
This patch implements the MDEV-31894 proposed fix to optimize the
workers_idle() check to compare the last in-use relay log’s
queued_count==dequeued_count for idleness. This removes the need for
workers_idle() to grab LOCK_parallel_entry, as these values are
atomically updated.
Huge thanks to Kristian Nielsen for diagnosing the problem!
Reviewed By:
============
Kristian Nielsen <knielsen@knielsen-hq.org>
Andrei Elkin <andrei.elkin@mariadb.com>
Add a test case that demonstrates a working setup as described in MDEV-26632.
This requires --gtid-ignore-duplicates=1 and --gtid-strict-mode=0.
In A->B->C, B filters some (but not all) events from A. C is promoted to
create A->C->B, and the current GTID position in B contains a GTID from A that
is not present in C (due to filtering). Demonstrate that B can still connect
with GTID to C, starting at the "hole" in the binlog stream on C originating
from A.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
Omit `state` when selecting processlist to verify which threads are running.
The state changes as threads are running (enter_state()), and this causes
sporadic test failures.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
Make sure the old binlog dump thread is not still running when manipulating
binlog files; otherwise there is a small chance it will see an invalid
partial file and report an I/O error.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
Let us remove a test that frequently fails with a result difference.
This test had been added in fc279d7ea2
to cover a bug in thd_destructor_proxy(), which was replaced with
simpler logic in 5e62b6a5e0 (MDEV-16264).
