Actions
Bug #16720
closedarv-mount exception handler bug
Story points:
-
Release:
Release relationship:
Auto
Description
Seen this a few times:
2020-08-17T18:00:02.957407737Z File "/usr/bin/arv-mount", line 10, in <module> 2020-08-17T18:00:02.957407737Z arvados_fuse.command.Mount(args).run() 2020-08-17T18:00:02.957407737Z File "/usr/share/python2.7/dist/python-arvados-fuse/lib/python2.7/site-packages/arvados_fuse/command.py", line 171, in run 2020-08-17T18:00:02.957491091Z recursive=self.args.unmount_all) 2020-08-17T18:00:02.957491091Z File "/usr/share/python2.7/dist/python-arvados-fuse/lib/python2.7/site-packages/arvados_fuse/unmount.py", line 161, in unmount 2020-08-17T18:00:02.957491091Z sys.stderr.write(fusermount_output) 2020-08-17T18:00:02.957491091Z NameError: global name 'sys' is not defined
Updated by Tom Clegg over 4 years ago
- Status changed from New to In Progress
16720-arv-mount-exception @ a46ef7496b83b2778de8db36e4948b55dddf3754 -- developer-run-tests: #2046
Updated by Lucas Di Pentima over 4 years ago
I think there's some code path that isn't properly exercised on our test suite, and that's the reason we didn't catch it earlier. Is it worth checking?
If not, this LGTM as it is.
Updated by Tom Clegg over 4 years ago
Yeah, this does indicate a testing gap. But this code path is hit only when we do "fusermount -u -z", check /proc/self/mountinfo, call realpath(), check /proc/self/mountinfo again, and the mount is still listed. Making this happen in a test seems too painful right now...
Updated by Anonymous over 4 years ago
- Status changed from In Progress to Resolved
Applied in changeset arvados|ffe94bef9cd17abb522d7fabb32326405d466a94.
Actions