BugComment:EntryResource timeouts

[Robert Collins]

  3 SELECT BugTask.assignee, BugTask.bug, BugTask.bugwatch, BugTask.date_assigned, BugTask.date_close ... roduct, BugTask.productseries, BugTask.distroseries, BugTask.milestone, BugTask.sourcepackagename:
   GET: 3 Robots: 0 Local: 1
      1 https://bugs.launchpad.net/api/devel/ubuntu/+source/compiz/+bug/710377/comments/4 (BugComment:EntryResource)
       OOPS-1950DQ231

20. 60 14004ms SQL-launchpad-main-master
SELECT BugTask.assignee,
       BugTask.bug,
       BugTask.bugwatch,
       BugTask.date_assigned,
       BugTask.date_closed,
       BugTask.date_confirmed,
       BugTask.date_fix_committed,
       BugTask.date_fix_released,
       BugTask.date_incomplete,
       BugTask.date_inprogress,
       BugTask.date_left_closed,
       BugTask.date_left_new,
       BugTask.date_triaged,
       BugTask.datecreated,
       BugTask.distribution,
       BugTask.distroseries,
       BugTask.heat,
       BugTask.id,
       BugTask.importance,
       BugTask.milestone,
       BugTask.OWNER, BugTask.product,
                      BugTask.productseries,
                      BugTask.sourcepackagename,
                      BugTask.status,
                      BugTask.statusexplanation,
                      BugTask.targetnamecache
FROM BugTask
WHERE BugTask.bug = 710377
ORDER BY BugTask.distribution,
         BugTask.product,
         BugTask.productseries,
         BugTask.distroseries,
         BugTask.milestone,
         BugTask.sourcepackagename

Seems inexplicably slow (only one task on the bug)

[Stuart Bishop (stub)]

This is from qa staging the day before a rollout, so likely someone doing database updates. Suggest we ignore this unless we see it on production or staging during more normal times.

[Robert Collins (lifeless)]

(last seen was qastaging, this was on prod)

[Stuart Bishop (stub)]

Misread the report. Its from production.

[Stuart Bishop (stub)]

There were several appserver requests cancelled at the same time this OOPS report was generated, all blocked trying to query the BugTask table. One of these must have been holding a lock. These will have all generated an OOPS at 2011-05-04 11:18:48

[Stuart Bishop (stub)]

None of the queries look unusual, so I can't see any way of tracing this further from the PG logs. The OOPS reports might be more insightful if we can track them down.

[William Grant (wgrant)]

The OOPSes are long gone and I haven't seen this happen since.