-
Notifications
You must be signed in to change notification settings - Fork 4.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make WriteBuffer delete all its temp files in /dev/shm #40652
Conversation
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-40652/33974
|
A new Pull Request was created by @wddgit (W. David Dagenhart) for master. It involves the following packages:
@cmsbuild, @smuzaffar, @Dr15Jones, @makortel can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
please test |
Looks good. Thanks! |
+1 Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-540120/30289/summary.html Comparison SummarySummary:
|
+1 |
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @rappoccio (and backports should be raised in the release meeting by the corresponding L2) |
+1 |
PR description:
Make WriteBuffer delete all its temp files in /dev/shm.
Each time the unit tests run the bug fixed here leaves behind 4 files with names similar to the following.
The 5 digit number is the PID. There is a possibility of a name clash if a later unit test both has the same PID AND is run by a different user. I noticed this while testing an unrelated PR and probably unluckily hit such a name clash.
This is related to #40484. Note that on cmsdev32 there are a few thousand old files in /dev/shm related to this problem cluttering up that directory. Some are there from problems fixed in #40484 and some from problems fixed by this PR. There could be some on other machines as well. I'm not sure how the machines are configured, but I would guess these will cleaned up next time each machine is rebooted. Until then, there is a possibility of name clashes in these unit tests (probably rarely). We believe the IBs avoid the problem because the code does try to delete the clashing file if it can. That does not work if a different user owns the file though, so for users running the unit tests outside the IBs there is a potential for problems.
PR validation:
Existing unit tests pass. I manually ran them and manually examined the contents of /dev/shm and new files are not left there anymore when the unit tests run. They appear to be getting cleaned up when the test ends as intended.