Re: [SQU] Out of inodes?

On Tue, 31 Oct 2000, Raul Alvarez Venegas wrote:

>
> Hi!
>
> I've been running Squid 2.4DEVEL4 for several weeks without problems but
> today stopped working and the following error in cache.log:
>
> 2000/10/31 21:08:55| Starting Squid Cache version 2.4.DEVEL4 for
> sparc-sun-solaris2.7...

> 2000/10/31 21:08:56| diskHandleWrite: FD 8: disk write error: (28) No
> space left on device

> # df -F ufs -o i
> Filesystem iused ifree %iused Mounted on
> ...
> /dev/dsk/c0t0d0s3 420783 82065 84% /cache1
> /dev/dsk/c0t0d0s4 403655 99193 80% /cache2
> /dev/dsk/c0t0d0s5 398054 104794 79% /cache3
>
> And the % of usage:
>
> # df -k
> Filesystem kbytes used avail capacity Mounted on
> ...
> /dev/dsk/c0t0d0s3 4131866 3210859 879689 79% /cache1
> /dev/dsk/c0t0d0s4 4131866 3213023 877525 79% /cache2
> /dev/dsk/c0t0d0s5 4131866 3211308 879240 79% /cache3

I think there are a number of possible bugs at play here.

1) solaris default filesystem has some weird problems with squid

    http://www.squid-cache.org/Doc/FAQ/FAQ-14.html#ss14.1
    (see the last subsection)

2) there was a bug that causes too many removals on the first
   cache dir, and not enough on the last cache dir. If you
   use truncate instead of unlink it could cause you to run
   out of inodes.

    http://www.squid-cache.org/Versions/v2/2.4/bugs/#squid-2.4.devel4-cachedir_imbalance

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html