Bug 477001

Summary: ext4: "No space left on device" incorrectly reported
Product: [Fedora] Fedora Reporter: Evan McNabb <emcnabb>
Component: kernelAssignee: Eric Sandeen <esandeen>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: high    
Version: 10CC: kernel-maint, quintela
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-03-12 03:02:15 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
tune2fs output none

Description Evan McNabb 2008-12-18 15:42:38 UTC
Description of problem:
The root filesystem on my Fedora 10 laptop is ext4. Multiple times now (every few days it seems like) the filesystem reports back there is no space left on the device, even though that's not the case. If I reboot the filesystem goes back to a normal state:

# touch /testfile.txt
touch: cannot touch `/testfile.txt': No space left on device
# mount
/dev/mapper/vg01-lv01 on / type ext4 (rw)
...
# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/mapper/vg01-lv01 15094792   4865328   9462696  34% /
...
# df -i
Filesystem            Inodes   IUsed   IFree IUse% Mounted on
/dev/mapper/vg01-lv01 958464  192170  766294   21% /
...

I do not see any errors in dmesg that are related to this.

Version-Release number of selected component (if applicable):
kernel-2.6.27.7-134.fc10.i686

How reproducible:

I have not noticed any usage pattern that triggers this. In my most recent case the laptop had been on for only a few minutes.

There is a suspend to RAM bug I'm fighting right now that often makes me hard poweroff my laptop, so that could possibly be a factor (although the filesystem always comes up fine after boot).

Additional info:

I'm attaching the tune2fs output at the time of occurrence.

Comment 1 Evan McNabb 2008-12-22 22:02:57 UTC
Created attachment 327701 [details]
tune2fs output

Comment 2 Eric Sandeen 2009-01-05 20:26:06 UTC
Evan, thanks, there are some known space-accounting problems yet... though also, I wonder about a patch I had in there causing problems.  The patch got inadvertently dropped, though, so we have an accidental chance to check something.  :)  Do you see the problem with, say, 2.6.27.8-140 and later?

Comment 3 Evan McNabb 2009-01-05 21:56:12 UTC
I upgraded to 2.6.27.9-159.fc10 on Dec 26th and I don't remember experiencing this since then. Definitely not in the last week. I'll keep watching for it to occur and will let you know if it does.

Comment 4 Evan McNabb 2009-01-22 18:51:31 UTC
Hi Eric,
I haven't seen this problem since upgrading to 2.6.27.9-159.fc10 which was almost a month ago. I'm fine with closing this bug if you'd like.

Comment 5 Eric Sandeen 2009-03-12 03:02:15 UTC
There was a bug in the inode allocator that was the most likely cause of this; it was fixed as of 2.6.29-0.122.rc5; I'm going to close on that assumption.

Thanks for the report, and sorry it took a while to track down :)