[MDEV-24315] Mariadb crash schema_table_store_record / run_fill_key_cache_tables Created: 2020-12-01  Updated: 2020-12-08  Resolved: 2020-12-08

Status: Closed
Project: MariaDB Server
Component/s: Information Schema, Plugin - feedback
Affects Version/s: 10.5.8
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Juan Gabriel Covas Assignee: Sergei Golubchik
Resolution: Duplicate Votes: 0
Labels: None
Environment:

CentOS 7


Issue Links:
Duplicate
is duplicated by MDEV-24279 Segfault after 1 day and 5 minutes up... Closed

 Description   

Mariadb crash, daily, consistently, after upgrade from 10.5.7 to 10.5.8. Simple database, just a lot of InnoDB tables and a few MyISAM tables. No stored procedures.

This is seen consistently at least once a day. The mariadb server crashes, gets restarted. The backtrace is consistent too about json_locate_key, fill_variables... etc.

201128 14:18:43 [ERROR] mysqld got signal 11 ;
[...]
201129 14:24:08 [ERROR] mysqld got signal 11 ;
[...]
201130 14:29:28 [ERROR] mysqld got signal 11 ;
[...]
Server version: 10.5.8-MariaDB-log
key_buffer_size=134217728
read_buffer_size=134217728
max_used_connections=1
max_threads=514
thread_count=15
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 134886105 K  bytes of memory
[...]
Thread pointer: 0x7f9ba80eb4a8
[...]
stack_bottom = 0x7f9bc4ff7f30 thread_stack 0x49000
??:0(my_print_stacktrace)[0x5555624ba37e]
??:0(handle_fatal_signal)[0x555561ec8fe7]
sigaction.c:0(__restore_rt)[0x7f9d36f12630]
??:0(wsrep_notify_status(wsrep::server_state::state, wsrep::view const*))[0x55556219f537]
??:0(wsrep_notify_status(wsrep::server_state::state, wsrep::view const*))[0x55556219a8bc]
??:0(schema_table_store_record(THD*, TABLE*))[0x555561d336c0]
??:0(run_fill_key_cache_tables(char const*, st_key_cache*, void*))[0x555561d3a32c]
??:0(fill_variables(THD*, TABLE_LIST*, Item*))[0x555561d3b486]
??:0(json_locate_key)[0x555562523258]
/usr/sbin/mariadbd(+0x6527a0)[0x555561bab7a0]
??:0(json_locate_key)[0x55556252337a]
pthread_create.c:0(start_thread)[0x7f9d36f0aea5]
??:0(__clone)[0x7f9d3503396d]
[...]
Query (0x0): (null)
Connection ID (thread ID): 1
Status: NOT_KILLED

Please let me know if I can provide more information on this. Cannot identify what's causing it.



 Comments   
Comment by Juan Gabriel Covas [ 2020-12-01 ]

Seems the same as MDEV-24279

Comment by Juan Gabriel Covas [ 2020-12-02 ]

It doesn't crash anymore once disabled the feedback plugin via config (feedback=OFF) and restarted mariadb, as advised in MDEV-24279

Generated at Thu Feb 08 09:29:03 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.