[MDEV-21096] Galera cluster crashed after starting async replication from master standlaone server Created: 2019-11-20  Updated: 2020-08-25  Resolved: 2019-11-26

Status: Closed
Project: MariaDB Server
Component/s: Galera
Affects Version/s: 10.4.8
Fix Version/s: 10.4.11

Type: Bug Priority: Major
Reporter: Nilnandan Joshi Assignee: Jan Lindström (Inactive)
Resolution: Fixed Votes: 1
Labels: None

Issue Links:
Relates
relates to MDEV-19572 Async slave on Galera node stops due ... Closed

 Description   

Environment is A (stand alone master) - > (B<>C,<>D) Galera
After upgrade to 10.4.8 and trying to start replication, server is getting crashed

2019-11-18 7:23:32 1241 [Note] Slave I/O thread: Start asynchronous replication to master '****@****.com:3306' in log 'mysql-bin.000044' at position 829636879
2019-11-18 7:23:32 1241 [Note] Slave I/O thread: connected to master 'repl_usr@****:3306',replication started in log 'mysql-bin.000044' at position 829636879
2019-11-18 7:23:32 1242 [Note] Slave SQL thread initialized, starting replication in log 'mysql-bin.000044' at position 829636879, relay log './relay-bin.000001' position: 4
mysqld: /home/buildbot/buildbot/padding_for_CPACK_RPM_BUILD_SOURCE_DIRS_PREFIX/mariadb-10.4.8/wsrep-lib/include/wsrep/client_state.hpp:314: int wsrep::client_state::app
end_key(const wsrep::key&): Assertion `state_ == s_exec' failed.
191118 7:23:33 [ERROR] mysqld got signal 6 ;
...
mysqld(my_print_stacktrace+0x2e)[0x5601fc7b183e]
mysqld(handle_fatal_signal+0x30f)[0x5601fc247e0f]
sigaction.c:0(__restore_rt)[0x7f20028e05d0]
:0(__GI_raise)[0x7f2000bb22c7]
:0(__GI_abort)[0x7f2000bb39b8]
:0(__assert_fail_base)[0x7f2000bab0e6]
:0(__GI___assert_fail)[0x7f2000bab192]
addr2line: 'mysqld': No such file
mysqld(wsrep_thd_append_key+0x196)[0x5601fc1c3f96]
mysqld(+0xa40103)[0x5601fc3e3103]
mysqld(+0xa4ce65)[0x5601fc3efe65]
mysqld(+0xa5ccd0)[0x5601fc3ffcd0]
mysqld(_ZN7handler12ha_write_rowEPKh+0x2f8)[0x5601fc253b38]
mysqld(_ZN15rpl_slave_state11record_gtidEP3THDPK8rpl_gtidybbPPv+0x381)[0x5601fc1687a1]
mysqld(_ZN15rpl_slave_state4loadEP3THDPKcmbb+0x9b)[0x5601fc169afb]
mysqld(+0x5f1234)[0x5601fbf94234]
mysqld(handle_slave_io+0xc62)[0x5601fbf9d602]
mysqld(+0xdc124d)[0x5601fc76424d]
pthread_create.c:0(start_thread)[0x7f20028d8dd5]
/lib64/libc.so.6(clone+0x6d)[0x7f2000c7a02d]
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort.
Query (0x0): is an invalid pointer
Connection ID (thread ID): 62
Status: NOT_KILLED

Binlog content at this position is

Log_name | Pos | Event_type | Server_id | End_log_pos | Info
| mysql-bin.000044 | 829636879 | Gtid | 4 | 829636921 | BEGIN GTID 140-4-391018 |
| mysql-bin.000044 | 829636921 | Query | 4 | 829637071 | use `db`; delete from table where last_used <= '2019-11-16 22:47:25.275' |
| mysql-bin.000044 | 829637071 | Query | 4 | 829637153 | COMMIT |



 Comments   
Comment by Seppo Jaakola [ 2019-11-20 ]

Please share the configuration of mariadb replication master server and galera slave node

Comment by Seppo Jaakola [ 2019-11-20 ]

I can reproduce close to matching crash of galera slave node, when using configuration:
gtid_pos_auto_engines=InnoDB

but, otoh, the crash pattern is not exactly similar

Generated at Thu Feb 08 09:04:33 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.