493170 - Intermittent failure of reftest 427129-table-caption.html: timed out waiting for onload to fire

Status: RESOLVED INVALID

(Release Engineering :: General, defect, P3)

Description

[(no longer active)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.5/1242337479.1242345375.18224.gz
Linux mozilla-1.9.1 unit test on 2009/05/14 14:44:39

This is intermittent because the only changeset in the range from the previous test cycle of this machine <http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.5/1242333381.1242341038.11991.gz> (which passed the reftest) is an OS X only patch <http://hg.mozilla.org/releases/mozilla-1.9.1/rev/d5a1a75a42d7>.

Comment 1

[Bernd]

I would say machine problem rather than test case issue

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

This certainly isn't a tables bug; it's either a reftest harness bug or a build machine issue.  I suspect the latter; if the machines are so overloaded sometimes that we can't load a 28K HTML file in 10 seconds, we've got a problem.

Comment 3

[Chris AtLee [:catlee]]

The reftest step started at 15:43:48.

At 15:48:09, VMWare initiated a migration off of bm-vmware07 to bm-vmware08.
At 15:48:37 the migration was completed.

At 16:04:28 the reftest step completed.

mrz, is it possible that the virtual machine was suspended for more than 10 seconds while being migrated?  Or should the migration be transparent to the VM?

Comment 4

[Chris AtLee [:catlee]]

moz2-linux-slave18 FTR.

Comment 5

[[e-mail] (:bhearsum)]

For the record, here's what the load on our VM hosts and storage arrays looked like shortly after this bug was moved to RelEng:
CPU Usage: 23%, 45%, 70%, 77%, 53%, 56%, 47%, 16%, 58%, 39%, 35%, 42%, 30%
RAM Usage: 46%, 41%, 42%, 42%, 60%, 79%, 63%, 59%, 65%, 69%, 70%, 67%, 74%

I'm trying to get the storage latency, but Nagios is set to off if the latency is over 20ms - so I'm sure it's in a fine state.

The migration that Catlee points out is quite suspicious, though.

Comment 6

[Lukas Blakk [:lsblakk] use ?needinfo]

Let's move this to future in case it happens again.  The machine migration looks the most suspicious here.  If it re-occurs please add info to this bug so we can determine how bad this is.

Comment 7

[Chris Cooper [:coop] (he/him)]

Mass move of bugs from Release Engineering:Future -> Release Engineering. See
http://coop.deadsquid.com/2010/02/kiss-the-future-goodbye/ for more details.

Comment 8

[Nick Thomas [:nthomas] (UTC+12)]

No recent reports, -> INVALID.