[Bug 89621] EXT4-fs error (device dm-1): ext4_mb_release_inode_pa:3773: group 24089, free 34, pa_free 32

[bugzilla-daemon@bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=89621

--- Comment #19 from John Hughes <john@calva.com> ---
(In reply to Theodore Tso from comment #17)
> John, are you willing to give us a lot more details about your set up?  
> Hardware, software, details of the host OS, details of the guest OS, what is
> the workload where the problem is happening, etc?
> 
> And it might be helpful if you opened a separate bug which we use
> exclusively for tracking your particular issue.

Ok, I've opened https://bugzilla.kernel.org/show_bug.cgi?id=102731

Please let me know if you need more info.

> One thing to keep in mind when we are talking about problems found in the
> guest OS is that misconfigurations at the host OS level can be
> indistinguishable from hardware problems.   So for example, depending on the
> caching policy used by KVM in the host OS issues, it may be that cache flush
> commands in the guest OS don't peculate down all the way down to the actual
> hardware platters.   Thus, after a power fail event, even if everything is
> configured correctly at the guest OS, it can result in an inconsistent file
> system image because the KVM / the host OS didn't make sure everything got
> pushed to the disk after a journal commit.

I have had one power fail, but since then mdadm has run a consistency check and
found no inconsistencies.

I'm also running regular (once a week) fsck's on (snapshots of) my filesystems.
 So far the fsck have never found a problem.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.