Bug #42842
closedCephFS linux kernel hang, v4.15
0%
Description
Simple file system operations like df and ls hang and show a status of D+ when running ps. dmesg logs sometimes show "cache_from_obj: Wrong slab cache. inode_cache but object is from ceph_inode_info". Have also seen ls segfault. Also when running fio to benchmark, I have seen "[517501.995780] watchdog: BUG: soft lockup - CPU#23 stuck for 23s! [fio:2110843]".
Possibly related to https://tracker.ceph.com/issues/42707
Seen on:
Linux - 4.15.0-70-generic #79~16.04.1-Ubuntu SMP Tue Nov 12 14:01:10 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
Linux - 4.15.0-66-generic #75~16.04.1-Ubuntu SMP Tue Oct 1 14:01:08 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
I have been testing version 4.15.0-46-generic, and so far it seems stable but I'm not confident yet.
Ceph version installed is "ceph version 14.2.4 (75f4de193b3ea58512f204623e6c5a16e6c1e1ba) nautilus (stable)"
Updated by Jeff Layton over 4 years ago
-66.75 is definitely bad, but -70.79 should be ok. Can you validate that you still see the problem on that kernel?
Updated by Adam Ludvik over 4 years ago
I am no longer seeing the problem on -70.79. Had a number of kernel versions installed and must have gotten confused.
Updated by Jeff Layton over 4 years ago
- Status changed from New to Resolved
Glad to hear it. We'll call this one resolved.