[LU-16982] Crash lustre after umount -d -f /mnt/lustre-mds4

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.16.0
• Affects Version/s: Lustre 2.16.0
• Severity: 3
• Assignee: Vitaliy Kuznetsov
• Reporter: Vitaliy Kuznetsov

Description

Crash seen during unmount while testing patch https://review.whamcloud.com/51731 "LU-16973 osd: adds SB_KERNMOUNT flag".

[15020.698384] Lustre: DEBUG MARKER: umount -d -f /mnt/lustre-mds4
[15021.262942] kernel BUG at fs/jbd2/transaction.c:378!
[15021.264875] CPU: 0 PID: 310834 Comm: kworker/0:2 4.18.0-477.15.1.el8_lustre.x86_64
[15021.267209] Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011
[15021.268314] Workqueue: events flush_stashed_stats_work [ldiskfs]
[15021.269610] RIP: 0010:start_this_handle+0x22c/0x520 [jbd2]
[15021.286052] Call Trace:
[15021.288463] jbd2__journal_start+0xee/0x1f0 [jbd2]
[15021.289403] jbd2_journal_start+0x19/0x20 [jbd2]
[15021.290310] flush_stashed_stats_work+0x36/0x90 [ldiskfs]
[15021.291398] process_one_work+0x1a7/0x360
[15021.293028] worker_thread+0x30/0x390
[15021.294554] kthread+0x134/0x150

 

Issue Links

• is duplicated by: LU-17017 sanity test_133f: server crashes in start_this_handle in jbd2
• is related to: LU-16973 Busy device after successful umount
• is related to: LU-16298 periodically write ldiskfs superblock