Page MenuHomePhabricator

db1069 s7 replication thread stuck on fawiki.flaggedrevs_tracking
Closed, ResolvedPublic

Description

Creating this ticket for the record.
The s7 instance was totally stuck and the instance was delayed 14h on this query:

|     17 | system user     |                   | fawiki             | Connect     |   53781 | updating                                                              | DELETE /* FlaggedRevs::clearStableOnlyDeps  */ FROM `flaggedrevs_tracking` WHERE ftr_from = '0' |    0.000 |

I was not even able to stop the slave.

I had to kill -9 the process.
Converting the table to InnoDB fixed the issue.

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptApr 18 2017, 6:12 AM
Marostegui closed this task as Resolved.Apr 18 2017, 6:12 AM
Marostegui moved this task from Triage to In progress on the DBA board.

This looks similar to: T161781

Marostegui updated the task description. (Show Details)Apr 18 2017, 6:15 AM

Mentioned in SAL (#wikimedia-operations) [2017-04-18T06:16:51Z] <marostegui> For the record: restarted s7 instance on db1069 - T163183