binlog_space_limit crashes PXC to during wsrep_recover mode

Description

When using binlog_space_limit PXC will not be able to perform wsrep recovery and crash.
Example:
1. set this in my.cnf
...
binlog_space_limit=5120M
...
2. attempt to start PXC node after crash - it will perform wsrep_recover and will crash:

{{02:19:00 UTC - mysqld got signal 11 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
Attempting to collect some information that could help diagnose the problem.
As this is a crash and something is definitely wrong, the information
collection process might fail.
Please help us make Percona XtraDB Cluster better by reporting any
bugs at https://jira.percona.com/projects/PXC/issues

key_buffer_size=67108864
read_buffer_size=524288
max_used_connections=0
max_threads=1001
thread_count=0
connection_count=0
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 848261 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x0
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0 thread_stack 0x40000
/usr/sbin/mysqld(my_print_stacktrace+0x2c)[0xed82fc]
/usr/sbin/mysqld(handle_fatal_signal+0x479)[0x7a8609]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xf890)[0x7f13eaf21890]
/usr/sbin/mysqld[0xe6b91b]
/usr/sbin/mysqld[0x7a0551]
/usr/sbin/mysqld(_Z11mysqld_mainiPPc+0xc62)[0x7a16c2]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f13e8e92b45]
/usr/sbin/mysqld[0x794f14]}}

Environment

None

Smart Checklist

Activity

Show:

Krunal Bauskar March 28, 2019 at 11:03 AM

commit 27f6769ac470ead5850fc6d1748c9e7ac9a3ddc2
Author: Krunal Bauskar <krunal.bauskar@percona.com>
Date: Wed Mar 27 10:55:05 2019 +0530

  • PXC#2490: binlog_space_limit crashes PXC to during wsrep_recover mode

  • As part of the "PXC#897: gitd_executed is empty post wsrep_recover"
    fix, PXC skips creation of new binlog file while running in recovery mode.

  • In light of this flow, PXC should also avoud purge action on binlog.

  • Patch adds the check for the same.

 

Krunal Bauskar March 27, 2019 at 5:02 AM

Thanks for the report. PXC during recovery avoids creating a new binlog file so when the purge operation is done binlog object doesn't have a valid new binlog file open. We will fix it to avoid running purge operation too when running PXC in recovery mode.

Done

Details

Assignee

Reporter

Time tracking

2h 45m logged

Priority

Smart Checklist

Created March 27, 2019 at 2:53 AM
Updated March 6, 2024 at 10:13 PM
Resolved March 28, 2019 at 11:03 AM