In article <Pine.NEB.4.61.0607061448210.16562@ugly.precedence.co.uk>,
Stephen Borrill  <netbsd@precedence.co.uk> wrote:
>On Thu, 6 Jul 2006, Quentin Garnier wrote:
>> On Thu, Jul 06, 2006 at 01:39:13PM +0000, Christos Zoulas wrote:
>>> In article <Pine.NEB.4.61.0607061102310.16562@ugly.precedence.co.uk>,
>>> Stephen Borrill  <netbsd@precedence.co.uk> wrote:
>>>> A customer's server has run out of inodes, but has a tiny amount of files:
>>>>
>>>> mailserver 3# find /usr | wc -l
>>>>   112200
>>>> mailserver 4# df -i /usr
>>>> Filesystem  1K-blocks      Used     Avail Capacity  iused    ifree  %iused
>>>> /dev/raid2a 153018716  17104340 128263444    11%  9539859    17131    99%
>>>>
>>>> That's nearly 100 times as many inodes as files/directories! Now, I know
>>>> an inode != file/directory, but it should be the same order of magnitude!
>>>
>>> Is it corrupted?
>
>That's all I can think.
>
>>> Or is there a process holding open tons of files?
>
>No, this is immediately after a reboot.
>
>> I remember having similar trouble on a very large file-system.  I think
>> it was corrupted.  I couldn't investigate, so I just newfs'd it after
>> moving away the data.
>
>It's a production server that I only have remote ssh access to, so I'd 
>rather avoid doing that (but it is possible: I could break the RAID1 array 
>and repartition one HD to hold the data temporarily).
>
>Might a forced fsck from single-user help (if I can talk them through 
>it!)?

That is what I would do first. Or even an fsck -n in multi-user mode to
see whats going on.

christos