when a slave stops, pt-table-checksum process ends

[Miguel Angel Nieto]

We have a master with multiple slaves. While the checksum process is ongoing if we stop a slave server the checksum process is also stopped:

ExitStatus:111 04-14T03:00:02 DBI connect(';host=x.x.x.x.x;mysql_read_default_group=client','xxxxx',...) failed: Can't connect to MySQL server on 'x.x.x.x' (111) at /usr/local/bin/pt-table-checksum line 1586

If the tool is disconnected from a slave it should just wait there until it reconnects again with a good timeout like 1 hour. That timeout could be configured with a new variable. The tool should only wait if the error is "Can't connect to MySQL server" but not when the error is related with user privileges (those are not going to be fixed automatically).

Comments are welcome :)

[Nilnandan Joshi (nilnandan-joshi)]

Confirmed. While running pt-table-checksum, if slave server will be stopped. It gives error. Timeout variable for this can be helpful.

nilnandan@desktop:~$ pt-table-checksum --user=root --password=msandbox --socket=/tmp/mysql_sandbox20886.sock --recursion-method dsn=h=localhost,D=percona,t=dsns --no-check-binlog-format --check-interval=10 TS ERRORS DIFFS ROWS CHUNKS SKIPPED TIME TABLE
04-28T12:19:57 0 0 0 1 0 0.006 mysql.columns_priv
04-28T12:19:57 0 0 0 1 0 0.004 mysql.db
04-28T12:19:57 0 0 0 1 0 0.005 mysql.event
04-28T12:19:57 0 0 0 1 0 0.004 mysql.func
04-28T12:19:57 0 0 40 1 0 0.005 mysql.help_category
04-28T12:19:57 0 0 485 1 0 0.257 mysql.help_keyword
04-28T12:19:58 0 0 1090 1 0 0.258 mysql.help_relation
04-28T12:19:58 0 0 533 1 0 0.263 mysql.help_topic
04-28T12:19:58 0 0 0 1 0 0.006 mysql.ndb_binlog_index
04-28T12:19:58 0 1 1 1 0 0.004 mysql.plugin
04-28T12:19:58 0 0 0 1 0 0.006 mysql.proc
04-28T12:19:58 0 0 0 1 0 0.003 mysql.procs_priv
04-28T12:19:58 0 0 2 1 0 0.006 mysql.proxies_priv
04-28T12:19:58 0 0 0 1 0 0.005 mysql.servers
04-28T12:19:58 0 0 0 1 0 0.005 mysql.tables_priv
04-28T12:19:58 0 0 0 1 0 0.006 mysql.time_zone
04-28T12:19:58 0 0 0 1 0 0.004 mysql.time_zone_leap_second
04-28T12:19:58 0 0 0 1 0 0.004 mysql.time_zone_name
04-28T12:19:58 0 0 0 1 0 0.005 mysql.time_zone_transition
04-28T12:19:58 0 0 0 1 0 0.006 mysql.time_zone_transition_type
04-28T12:19:58 0 0 8 1 0 0.006 mysql.user
04-28T12:19:58 0 1 1 1 0 0.006 percona.dsns
Replica desktop is stopped. Waiting.
04-28T12:20:16 Fatal error checksumming table test.nil: Lost connection to replica desktop while attempting to get its lag (DBI connect(';host=localhost;mysql_socket=/tmp/mysql_sandbox20887.sock;mysql_read_default_group=client','msandbox',...) failed: Can't connect to local MySQL server through socket '/tmp/mysql_sandbox20887.sock' (2) at /usr/bin/pt-table-checksum line 1576.) at /usr/bin/pt-table-checksum line 9695.

04-28T12:20:16 1 0 7570408 10 0 18.446 test.nil
nilnandan@desktop:~$

[Frank Cizmich (frank-cizmich)]

Marked as duplicate of:
https://bugs.launchpad.net/percona-toolkit/+bug/1042727