Closed Bug 1759084 Opened 3 years ago Closed 3 years ago

Perma systemsymbols tasks: Exception: The "common" site is out-of-date, even though it has already been activated. Was it modified while this Mach process was running?

Categories

(Toolkit :: Crash Reporting, defect)

defect

Tracking

()

RESOLVED FIXED
100 Branch
Tracking Status
firefox-esr91 --- unaffected
firefox98 --- unaffected
firefox99 --- unaffected
firefox100 --- fixed

People

(Reporter: NarcisB, Assigned: mhentges)

References

(Regression)

Details

(Keywords: intermittent-failure, regression)

Attachments

(3 files)

Push with failure:
https://treeherder.mozilla.org/jobs?repo=mozilla-central&resultStatus=testfailed%2Cbusted%2Cexception%2Cretry%2Cusercancel&fromchange=446c2fcf0c0b317aeca92bd4ee2182543e722c31&selectedTaskRun=Dn2DqBH0RuuvqRZXQdYSMA.0

Log link:
https://treeherder.mozilla.org/logviewer?job_id=370621931&repo=mozilla-central&lineNumber=159

Log snippet:

[task 2022-03-11T00:24:09.718Z] created virtual environment CPython3.6.9.final.0-64 in 23ms
[task 2022-03-11T00:24:09.718Z]   creator CPython3Posix(dest=/builds/worker/checkouts/gecko/obj-x86_64-pc-linux-gnu/_virtualenvs/common, clear=False, no_vcs_ignore=False, global=False)
[task 2022-03-11T00:24:09.718Z]   activators BashActivator,CShellActivator,FishActivator,PowerShellActivator,PythonActivator
[task 2022-03-11T00:24:10.023Z] Traceback (most recent call last):
[task 2022-03-11T00:24:10.023Z]   File "toolkit/crashreporter/tools/upload_symbols.py", line 260, in <module>
[task 2022-03-11T00:24:10.023Z]     sys.exit(main())
[task 2022-03-11T00:24:10.023Z]   File "toolkit/crashreporter/tools/upload_symbols.py", line 61, in main
[task 2022-03-11T00:24:10.023Z]     config.activate_virtualenv()
[task 2022-03-11T00:24:10.023Z]   File "/builds/worker/checkouts/gecko/python/mozbuild/mozbuild/base.py", line 863, in activate_virtualenv
[task 2022-03-11T00:24:10.023Z]     self.virtualenv_manager.activate()
[task 2022-03-11T00:24:10.023Z]   File "/builds/worker/checkouts/gecko/python/mach/mach/site.py", line 552, in activate
[task 2022-03-11T00:24:10.023Z]     self.ensure()
[task 2022-03-11T00:24:10.023Z]   File "/builds/worker/checkouts/gecko/python/mach/mach/site.py", line 530, in ensure
[task 2022-03-11T00:24:10.023Z]     f'The "{self._site_name}" site is out-of-date, even though it has '
[task 2022-03-11T00:24:10.023Z] Exception: The "common" site is out-of-date, even though it has already been activated. Was it modified while this Mach process was running?
[task 2022-03-11T00:24:10.053Z] Creating default state directory: /builds/worker/.mozbuild
[taskcluster 2022-03-11 00:24:10.396Z] === Task Finished ===
[taskcluster 2022-03-11 00:24:10.396Z] Unsuccessful task run with exit code: 1 completed in 153.653 seconds

Jon, could this be caused by Bug 1757956 ?

Flags: needinfo?(jcoppeard)
Flags: needinfo?(jcoppeard) → needinfo?(gsvelto)
Summary: Perma Tier2 Exception: The "common" site is out-of-date, even though it has already been activated. Was it modified while this Mach process was running? → Perma systemsymbols tasks: Exception: The "common" site is out-of-date, even though it has already been activated. Was it modified while this Mach process was running?

I highly doubt that this is a gsvelto-related issue :)
I'll dig in later today.

Flags: needinfo?(gsvelto)
Assignee: nobody → mhentges
Status: NEW → ASSIGNED
Flags: needinfo?(mhentges)

The Mach site management has been significantly simplified
in mach_initialize.py, so less of the MachSiteManager
API is used. Privatize such now-externally-unused functions,
and no longer return unneeded values.

There are cases in CI where sites are being considered out-of-date,
despite the situation being unexpected - for example, the
"site out-of-date even though already activated" issue described in
this patch's associated bug.

Return a reason string from _is_venv_up_to_date(), and print it in
cases where a virtualenv being out-of-date is erroneous.

Depends on D140854

(In reply to Mitchell Hentges [:mhentges] πŸ¦€ from comment #1)

I highly doubt that this is a gsvelto-related issue :)
I'll dig in later today.

Thanks! I'm in the process of making a significant overhaul of those tasks so I was super-puzzled that they broke even before I landed my changes :)

Attachment #9267407 - Attachment description: WIP: Bug 1759084: Tighten `MachSiteManager` API surface → Bug 1759084: Tighten `MachSiteManager` API surface
Attachment #9267408 - Attachment description: WIP: Bug 1759084: Print reason for site being out-of-date → Bug 1759084: Print reason for site being out-of-date

upload_symbols.py is always executed within a Mach site (otherwise
activating the common command site would fail), therefore
re-activating common is redundant.

Depends on D140855

Pushed by mhentges@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/13aef63ad6a5 Tighten `MachSiteManager` API surface r=ahal https://hg.mozilla.org/integration/autoland/rev/5f18914637a3 Print reason for site being out-of-date r=ahal https://hg.mozilla.org/integration/autoland/rev/412197c555c6 Don't redundantly activate site in upload_symbols.py r=ahal
Regressions: 1760742

Backed out for causing symbol bustages.

Status: RESOLVED → REOPENED
Flags: needinfo?(mhentges)
Resolution: FIXED → ---

Patch updated and re-landed.

Flags: needinfo?(mhentges)
Pushed by mhentges@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/3d254d42ed23 Tighten `MachSiteManager` API surface r=ahal https://hg.mozilla.org/integration/autoland/rev/c74d2b2bd792 Print reason for site being out-of-date r=ahal https://hg.mozilla.org/integration/autoland/rev/43f7556db2e1 Don't redundantly activate site in upload_symbols.py r=ahal
Component: Symbols → Crash Reporting
Product: Socorro → Toolkit
Target Milestone: --- → 100 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: