Bug 1432003
Summary: | After logrotate , dynamic looping call gets 'ValueError: I/O operation on closed file' on self.stream.flush() in /usr/lib64/python2.7/logging/handlers.py | ||
---|---|---|---|
Product: | Red Hat Enterprise Linux 7 | Reporter: | Masaki Furuta ( RH ) <mfuruta> |
Component: | python | Assignee: | Charalampos Stratakis <cstratak> |
Status: | CLOSED ERRATA | QA Contact: | Branislav Náter <bnater> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 7.0 | CC: | apevec, berrange, bnater, cstratak, dasmith, eglynn, jkejda, kchamart, lhh, mfuruta, pviktori, python-maint, rbryant, sbauza, sferdjao, sgordon, srevivo, torsava, vromanso, vstinner |
Target Milestone: | pre-dev-freeze | Keywords: | Triaged |
Target Release: | --- | Flags: | vstinner:
needinfo+
|
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | python-2.7.5-58.el7 | Doc Type: | No Doc Update |
Doc Text: |
undefined
|
Story Points: | --- |
Clone Of: | Environment: | ||
Last Closed: | 2017-08-01 20:25:03 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Comment 5
Victor Stinner
2017-03-31 14:33:37 UTC
Mike Burns confirmed that python-inotify is already in RHEL 7, so it is ok to use it in RHOS 6. (In reply to Petr Viktorin from comment #12) > Harris or Tomáš, can you please confirm that RHEL 7 uses 2.7.5 and that > https://hg.python.org/cpython/rev/bb8b0c7fefd0 fix was not backported? I can confirm that the python version in RHEL 7 (2.7.5), doesn't include this fix. Victor, by backporting the patch in python, the issue will be resolved? > Victor, by backporting the patch in python, the issue will be resolved?
Sorry, that's hard to say at this point:
"How reproducible: Once on customer site only."
and then "Actual results: Can't get any log entry to pursue the cause in nova-compute.log after Feb 05, due to 'ValueError: I/O operation on closed file' on self.stream.flush() in /usr/lib64/python2.7/logging/handlers.py, after logrotate."
Funny or not, I need missing logs to prove that the logging bugfix will fix the logging issue...
At least, according to the customer traceback and to the CPython upstream bug report: they seem *very* similar.
The real mystery is why the nova failed to reopen the nova file while logrotate was rotating the log file? Likely a race condition, likely hard to reproduce.
Scratch builds with the fix on top of python-2.7.5-39 https://brewweb.engineering.redhat.com/brew/taskinfo?taskID=13042090 Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2017:1868 |