New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 724533 link

Starred by 1 user
Status: Available
Owner: ----
Cc:
jochen@chromium.org
█ phajdan.jr@chromium.org
dpranke@chromium.org
iannu...@google.com
martiniss@chromium.org
jbudorick@chromium.org
scottmg@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Feature

Blocking:
issue 724334



Sign in to add a comment

"corrupt PDB file" errors on win_chromium_rel_ng; should clobber if compile (without patch) fails

Project Member Reported by kbr@chromium.org, May 19 2017 
There are several compile errors on win_chromium_rel_ng recently where a .pdb file became corrupt, causing the build to fail.

https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/?limit=200

This is especially visible by looking at "failed compile (with patch)". A few examples:

https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/449753
https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/449776
https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/449826
https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/449947

I'm not sure whether something changed recently to cause this incidence to increase (full disk? compiler toolchain upgrade breaking incremental linking sometimes?). Could someone please take a look at this? Maybe only one or a couple of builders are broken?

Not sure what mitigation should be done -- should we clobber the builder automatically if "compile (without patch)" fails?

Comment 1 by bpastene@chromium.org, May 22 2017

Owner: bpastene@chromium.org
Status: Assigned (was: Untriaged)
Looks like the bots mentioned above have recovered, but at least one other is suffering the same thing:
https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/451203

I'll take a look when I get a chance.

Comment 2 by bpastene@chromium.org, May 24 2017

Owner: ----
Status: Untriaged (was: Assigned)
Wasn't able to get to this during my shift. Releasing to the next trooper.

Comment 3 by jeffcarp@chromium.org, May 24 2017 

bpastene what were you going to take a look at? I'm new to troopering and trying to figure out how to diagnose.

Comment 4 by jbudorick@chromium.org, May 26 2017

Cc: dpranke@chromium.org jbudorick@chromium.org
Labels: -Infra-Troopers
I don't think this is a trooper issue any more, as we're beyond the short-term mitigation portion of this bug -- the last compile (without patch) failure due to a PDB issue was https://luci-milo.appspot.com/buildbot/tryserver.chromium.win/win_chromium_rel_ng/451663, on Monday.   Clobber on compile (without patch) failure would probably be a chromium recipe change.

Comment 5 by kbr@chromium.org, May 31 2017

Labels: -Pri-1 -Type-Bug Pri-2 Type-Feature
Agreed that clobbering the bot upon compile (without patch) failure would be a recipe change. Recategorizing as feature request for consideration.

While this could potentially increase cycle times in some situations (like a revert breaking the source tree), if we can detect from the logs that a situation like this one happened, we could clobber only in those situations.

Comment 6 by martiniss@chromium.org, Dec 13 2017

Status: Available (was: Untriaged)

Comment 7 by iannu...@google.com, Oct 19

Cc: -iannucci@chromium.org iannu...@google.com
►

Sign in to add a comment