Details
Description
There are a number of bug reports of cluster wide conflict resolving related crashes or hangs. To troubleshoot these, it would be good to run generic cluster conflict tests, just for easier reproducing of the issue(s).
Run high intensity write-write conflict load in the cluster, with varying SQL access patterns.
Also, some mtr tests exercising high priority victim aborting, appear to be disabled in 10.4. Enable these and fix the test if it appears to be non-deterministic.
Analyze and troubleshoot issues surfacing from the testing.
Attachments
Issue Links
- relates to
-
MDEV-30317 Transaction savepoint may cause failure in galera replaying
-
- Closed
-
Activity
This new RQG grammar makes two other scenarios to surface, I will submit separate MDEV issue for both, as they are not related to the fixes suggested in PR's for this MDEV. The surfacing issues are:
- KILL command is issued from inside of a transaction. And further, this KILL execution can be a target of another KILL command. BF aborting a KILL inside a transaction appears to cause a crash.
- SAVEPOINT command will be replicated as query log event. If transaction containing such SAVEPOINT event, has to replay, the replaying may fail for un-assigned character set information.
- bb-10.4-
MDEV-29684-galera- https://github.com/MariaDB/server/commit/c85f18a37feae677a1593fd7bbebd19bb4a5fd94 (sql)
- https://github.com/MariaDB/server/commit/3e466c745ef3a97e1372a23f37a0cf04745ce25f (test case)
- https://github.com/MariaDB/server/commit/c76134bb305e49689380e271edd959f656be4de0 (sql)
- https://github.com/MariaDB/server/commit/83a450464806625e8d77e1aeb93aa0f5b972ec59 (InnoDB)
- https://github.com/MariaDB/server/commit/54abb014a3abe2a2f09e7bdeb799b5614d9d94a4 (sql + InnoDB)
- bb-10.5-
MDEV-29684-galera (about same as 10.4) - bb-10.6-
MDEV-29684-galera- https://github.com/MariaDB/server/commit/bcf368f58883f7efc8676426c56853014d15857d
- https://github.com/MariaDB/server/commit/31a8b027422ad63e0b1bc91ed982a5ba309d7c51
- https://github.com/MariaDB/server/commit/2a1e41727cb9f4b563fb5bbefd4047d3e5878519
- https://github.com/MariaDB/server/commit/640b5d52a8d48934c980d7cb15c02661ad20e56d
- https://github.com/MariaDB/server/commit/2a92e74dd4611367c3581e1b6542c537863b4675
seppo, I provided some review comments to a 10.6 version of this: 1, 2, 3, 4. Please address it.
I did not look at other versions yet.
seppo We can reproduce the problem even without the {{SET wsrep_OSU_method..}}statement. This is a sporadic issue, so start the RQG run multiple times to reproduce the problem.
PFA grammar oltp_and_ddl_v2.yy
to reproduce the issue.
RQG command to reproduce the hang problem.
perl runall-new.pl \--basedir=/test/mtest/mariadb-10.4.28-linux-x86_64 \--vardir=/home/ramesh/galera-rqg-test \--mysqld=--wsrep-provider=/test/mtest/mariadb-10.4.28-linux-x86_64/lib/libgalera_smm.so \--gendata=conf/mariadb/oltp.zz \--grammar=conf/mariadb/oltp_and_ddl_v2.yy \--threads=16 \--galera=mss \--mysqld=--wsrep_sst_method=rsync \--mysqld=--core \--mysqld=--general-log \--mysqld=--general-log-file=queries.log \--mysqld=--log-output=file \--mysqld=--wsrep-debug=0 \--mysqld=--wsrep-sync-wait=15 \--mysqld=--wsrep_retry_autocommit=0 \--mysqld=--wsrep_slave_threads=12 \--mysqld=--wsrep_log_conflicts=1 \--mysqld=--wsrep_on=1 \--mysqld=--default-storage-engine=innodb \--mysqld=--sort_buffer_size=200M \--mysqld=--innodb-lock-wait-timeout=1 \--mysqld=--gtid_domain_id=10 \--mysqld=--wsrep_gtid_domain_id=100 \--mysqld=--wsrep_gtid_mode=ON \--mysqld=--wsrep_slave_threads=4 \--mysqld=--slave_parallel_threads=4 \--mysqld=--server_id=11 \--mysqld=--gtid_strict_mode=1 \--mysqld=--log_slave_updates=ON \--mysqld=--log_bin=binlog \--mysqld=--binlog_format=ROW \--mysqld=--master_info_repository=TABLE \--mysqld=--relay_log_info_repository=TABLE