Closed Bug 1420472 Opened 7 years ago Closed 6 years ago

Intermittent mozmake.EXE[4]: *** [../../../dist/xpi-stage/test_jar_mn/chrome/.mkdir.done] Error 1

Categories

(Firefox Build System :: General, defect, P5)

Product:
Firefox Build System
Component:
General
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED INCOMPLETE

People

(Reporter: intermittent-bug-filer, Unassigned)

References

Details

(Keywords: intermittent-failure, Whiteboard: [stockwell disabled]) 

What's quoted in the summary is bug 1416977 and is a red herring.

The real issue in that log is:
16:01:15     INFO - mkdir -p '../../../dist/xpi-stage/test_jar_mn/chrome/'
16:01:15     INFO - mkdir: cannot create directory `..': No such file or directory
16:01:15     INFO - z:/build/build/src/config/makefiles/autotargets.mk:59: recipe for target '../../../dist/xpi-stage/test_jar_mn/chrome/.mkdir.done' failed
16:01:15     INFO - mozmake.EXE[4]: *** [../../../dist/xpi-stage/test_jar_mn/chrome/.mkdir.done] Error 1
I am seeing this a lot in a try push I did for bug 1395933, to
enable execution of Rust unit tests on try:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=35d8193e484ceb6f995469c8d4258a535e97ba7a&group_state=expanded
(In reply to Andreas Tolfsen ‹:ato› from comment #5)
> I am seeing this a lot in a try push I did for bug 1395933, to
> enable execution of Rust unit tests on try:
> https://treeherder.mozilla.org/#/
> jobs?repo=try&revision=35d8193e484ceb6f995469c8d4258a535e97ba7a&group_state=e
> xpanded

See comment 1.
Build Bustage started to occur from this push: https://treeherder.mozilla.org/#/jobs?repo=autoland&revision=dfba797e7a1fb03d071e913e73a3f43d650b0bc3
on windows2012-64-rusttests opt
Andreas Tolfsen can you please take a look?
Flags: needinfo?(ato)
So I don’t think that https://bugzil.la/1395933 caused that
intermittent, but from the latest OrangeFactor update it does seem
like it may have increased the frequency of the intermittent somehow.
I don’t understand how adding a couple of more Rust unit tests will
have increased the frequency, so maybe it’s worth waiting for another
OrangeFactor update to see if this is really the case.
Flags: needinfo?(ato)
Depends on: 1425365
This is an perma failure now. It`s failing on every "windows2012-32-rusttests opt"

A recent log file:  
https://treeherder.mozilla.org/logviewer.html#?job_id=151723507&repo=mozilla-inbound

:ato can you please take a look?
Flags: needinfo?(ato)
(In reply to Narcis Beleuzu [:NarcisB] from comment #12)

> This is an perma failure now. It`s failing on every
> "windows2012-32-rusttests opt"

I disabled the geckodriver Rust unit tests on Windows in
https://bugzil.la/1425365, which should revert the intermittency
rate back to ~0.03.
Flags: needinfo?(ato)
(In reply to Mike Hommey [:glandium] from comment #1)
> What's quoted in the summary is bug 1416977 and is a red herring.
> 
> The real issue in that log is:
> 16:01:15     INFO - mkdir -p '../../../dist/xpi-stage/test_jar_mn/chrome/'
> 16:01:15     INFO - mkdir: cannot create directory `..': No such file or

Something is quoting the path there. Was there a recent change to the recursive make backend or emitter?
(In reply to Ted Mielczarek [:ted.mielczarek] from comment #14)
> (In reply to Mike Hommey [:glandium] from comment #1)
> > What's quoted in the summary is bug 1416977 and is a red herring.
> > 
> > The real issue in that log is:
> > 16:01:15     INFO - mkdir -p '../../../dist/xpi-stage/test_jar_mn/chrome/'
> > 16:01:15     INFO - mkdir: cannot create directory `..': No such file or
> 
> Something is quoting the path there. Was there a recent change to the
> recursive make backend or emitter?

The quotes shouldn't be causing a problem. In fact, the error message seems to indicate it's not causing a problem. But it fails to create "..", which suggests the current working directory somehow doesn't exist at the moment of the error, that is, it has been removed from under the shell's feet.

Now, interestingly, this has nothing to do with all the recent messages in this bug, thanks to bug 1416977 hiding/masking real different errors.
Whiteboard: [stockwell needswork]
Blocks: 1425365
No longer depends on: 1425365
See Also: → 1416977
possibly this is fixed or we don't run rusttests anymore?
Whiteboard: [stockwell disable-recommended] → [stockwell unknown]
(In reply to Joel Maher ( :jmaher) (UTC-5) from comment #22)
> possibly this is fixed or we don't run rusttests anymore?

See https://bugzilla.mozilla.org/show_bug.cgi?id=1420472#c13.
Whiteboard: [stockwell unknown] → [stockwell disabled]
Summary: Intermittent ValueError: I/O operation on closed file → Intermittent mozmake.EXE[4]: *** [../../../dist/xpi-stage/test_jar_mn/chrome/.mkdir.done] Error 1
Product: Core → Firefox Build System
This seems to have stopped happening, although it was very low frequency to begin with.
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.