Bug 1181322 - *eating filesystems* : ignores check count/time bypassing needed filesystem check
Summary: *eating filesystems* : ignores check count/time bypassing needed filesystem c...
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Fedora
Classification: Fedora
Component: systemd
Version: 21
Hardware: Unspecified
OS: Unspecified
unspecified
urgent
Target Milestone: ---
Assignee: systemd-maint
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2015-01-12 21:16 UTC by Richard Z.
Modified: 2015-01-14 01:20 UTC (History)
10 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2015-01-14 01:20:28 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Richard Z. 2015-01-12 21:16:21 UTC
after upgrade to F21 I was suspecting filesystem problems and thus have done 
   tune2fs -C 1 dev
for all filesystems.

After reboot filesystem checks are skipped, nothing checked. In the logs I see tons of messages like 

 EXT4-fs (dm-9): warning: checktime reached, running e2fsck is recommended

Comment 1 Michal Schmidt 2015-01-13 12:53:30 UTC
(In reply to Richard Z. from comment #0)
> after upgrade to F21 I was suspecting filesystem problems and thus have done 
>    tune2fs -C 1 dev

This does not look right. "-C" sets the mount count. Setting it to a low value seems counter-productive with respect to your stated goal.

Please attach your /etc/fstab and the output of "tune2fs -l $dev".

Comment 2 Richard Z. 2015-01-13 13:07:40 UTC
sorry for the noise. Seems like I got the wrong option in the hurry.

Maximum mount count:      -1

Seems ok to close this particular bug report.

I am wondering if I should open a separate one because apparently no full filesystem check was forced after these messages?


 kernel: JBD2: Spotted dirty metadata buffer (dev = dm-5, blocknr = 0). There's a risk of filesystem corruption in case of system crash.

Or is it expected that this kind of fs damage will be repaired without full fs checks?

Comment 3 Zbigniew Jędrzejewski-Szmek 2015-01-14 01:20:28 UTC
(BTW.: systemd does not determine whether a check should be performed based on the filesystem mount counts or time-since-last-check. It leaves the job to fsck, which in turn leaves the job to fsck.extX.)

This is a kernel message. You could file a bug against the kernel package, as they will be qualified to answer whether this is something that can be fixed with a filesystem check, something to ignore, or a sign of other problems.


Note You need to log in before you can comment on or make changes to this bug.