[LU-5729] VFS: Busy inodes after unmount of loop0. Self-destruct in 5 seconds. Have a nice day...

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.7.0
• Labels: patch
• Severity: 3
• Assignee: James Nunez (Inactive)
• Reporter: Sergey Cheremencev

Description

Extra inode decrement is needed for scrub->os_inode(OI_scrub) in osd_scrub_setup in case of ENOSPACE error.
It is simple to reproduce:

1. mount ost as ldiskfs to mntpt
2. fill mnt point to get ENOSPACE later(llverfs may be used for that)
3. umount ost
4. delete scrub files using writeconf
5. mount ost

These causes kernel message like below:

VFS: Busy inodes after unmount of loop0. Self-destruct in 5 seconds. Have a nice day...

xyratex-bug-id: MRP-2109

Issue Links

• is related to: LU-5986 conf-sanity test_83, test_84: failed to start OST
• is related to: LU-6207 conf-sanity test_83: test failed to respond and timed out
• is related to: LU-9433 sanity-scrub test_6: Error in dmesg detected
• is related to: LU-6468 Interop 2.5.3<->master conf-sanity test_83: No space left on device