Re: [SQU] The old cache over limit problem

On Thu, 16 Nov 2000, Ethy H. Brito wrote:

>
> Dear Squidere
>
> The thing is getting worse here! Almost no light at the end of the tunnel!
>
> servidor:/cache# df
> Filesystem 1k-blocks Used Available Use% Mounted on
> /dev/hda8 1011928 955084 5440 99% /cache
>
> Even after I have changed my replacement_policy from LFUDA to LRU. Squid
> refuses to delete files! (obviouly I restarted it!)

did you show us your 'client mgr:storedir' output?

does cache.log complain about being over the limit?

Is the unlinkd process running?

>
> I was invetigating a little and look what I found:
>
> servidor:/cache# du -ks *
> 549196 00
> 386644 01
> 1028 02
> 1028 03
 
> Shouldn't it be an even distribution of files over the directory tree?

no, see the FAQ.

if you want them spread out more, decrease the number of L2 directories.

Duane W.

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