search

dmwm / WMCore

Core workflow management components for CMS.
Apache License 2.0
46 stars 107 forks source link

sometimes voms-proxy-info fails with exitcode=0 #12083

Closed belforte closed 2 months ago

belforte commented 2 months ago

Impact of the bug CRAB TaskWorker

Describe the bug I know it is hard to believe, and I could not reproduce this in a test stand, so it is likely related to load or concurrency, but I have evidence that at times voms-proxy-info -file xxxx -timeleft was returning exit code = 0 while file xxxx is not there

How to reproduce it I could not reproduce it myself. But somehow it was persistently failing for 4 users (not others) until I patched the code.

Expected behavior Even if retcode is 0, should properly handle any exception from parsing the stdout

Additional context and error message see https://github.com/dmwm/CRABServer/issues/8625#issuecomment-2312553960

I will make a tested PR

belforte commented 2 months ago

@novicecpp correctly pointed out that this happened becasue of my change #12075 where I have changed command to <command> | tail -1 , so the exit code comes from the tail command.

I will address that as well.

belforte commented 2 months ago

I am closing this since it was a wrong diagnosis. Correct code will appear as a fix to https://github.com/dmwm/WMCore/issues/12075#issuecomment-2313015599

Sorry for the mess