[CentOS] Re: Strange Problem with dm-0
William L. Maltby
CentOS4Bill at triad.rr.com
Mon Jan 7 23:02:47 UTC 2008
On Mon, 2008-01-07 at 17:22 -0500, James B. Byrne wrote:
> # ls -l /dev/mapper /dev/dm*
> brw-r----- 1 root root 253, 0 Jan 7 16:42 /dev/dm-0
> brw-r----- 1 root root 253, 1 Jan 7 16:42 /dev/dm-1
> brw-r----- 1 root root 253, 2 Jan 7 16:42 /dev/dm-2
> brw-r----- 1 root root 253, 3 Jan 7 16:42 /dev/dm-3
> brw-r----- 1 root root 253, 4 Jan 7 16:42 /dev/dm-4
> brw-r----- 1 root root 253, 5 Jan 7 16:42 /dev/dm-5
> brw-r----- 1 root root 253, 6 Jan 7 16:42 /dev/dm-6
>
> /dev/mapper:
> total 0
> crw------- 1 root root 10, 63 Jan 7 16:42 control
> brw-rw---- 1 root disk 253, 0 Jan 7 16:42 VolGroup00-LogVol00
> brw-rw---- 1 root disk 253, 2 Jan 7 16:42 VolGroup00-LogVol01
> brw-rw---- 1 root disk 253, 1 Jan 7 16:42 VolGroup00-LogVol02
> brw-rw---- 1 root disk 253, 3 Jan 7 16:42 VolGroup00-lv--IMAP
> brw-rw---- 1 root disk 253, 6 Jan 7 16:42 VolGroup00-lv--IMAP--2
> brw-rw---- 1 root disk 253, 4 Jan 7 16:42 VolGroup00-lv--MailMan
> brw-rw---- 1 root disk 253, 5 Jan 7 16:42 VolGroup00-lv--webfax
>
> I infer that dm-0 ===> VolGroup00-LogVol00 and that
> VolGroup00-LogVol00 ===> /
>
> so df / gives
>
> # df /
> Filesystem 1K-blocks Used Available Use% Mounted on
> /dev/mapper/VolGroup00-LogVol00
> 8256952 6677880 1159644 86% /
>
>
> I am guessing that the yum update caused the file system to fill and to
> precipitate this problem. Is the the probable cause?
With appx 1.19GB available, by itself I don't think so. There are
underlying tmpfs files systems associated with the LVs. A stat on those
will show a different set of numbers. Maybe one of these filled?
# stat --filesystem /dev/mapper/VolGroup00-LogVol00
File: "/dev/mapper/VolGroup00-LogVol00"
ID: 0 Namelen: 255 Type: tmpfs
Blocks: Total: 194473 Free: 194415
Available: 194415 Size: 4096
Inodes: Total: 194473 Free: 194075
# stat --filesystem /dev/mapper/VolGroup01-Home01
File: "/dev/mapper/VolGroup01-Home01"
ID: 0 Namelen: 255 Type: tmpfs
Blocks: Total: 194473 Free: 194415
Available: 194415 Size: 4096
Inodes: Total: 194473 Free: 194075
# df -H /
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
19G 11G 7.4G 59% /
I *guess* that when doing the update, there were some components that
were in use and could not be *truly* deleted. When the update was being
done, in addition to temporary high-water marks achieved while
transactions were being done, rpms shuffled, etc., there was probably
additional space not yet released by some component that was "replaced"
but could not yet be deleted.
There is also the possibility that your i-nodes were used up. Since I
set my file systems to 4K blocks, I use fewer than normal.
Do "df -i" on mounted FSs.
I *suspect*, relative to your orphaned i-node, that this same underlying
situation was the cause? Some component that couldn't be released was
still active when the file system had to be unmounted. It should no re-
cur unless it really is some other problem.
--
Bill
More information about the CentOS
mailing list