Re: [SQU] cache_dir problem

If you're on 2.3.STABLE3, make sure you apply the patches for it, including
this one:
http://www.squid-cache.org/Versions/v2/2.3/bugs/#squid-2.3.stable3-storeExpiredReferenceAge

By cutting the cache size this close to the filesystem size, you may have
enough overhead to fill it anway. Directory structures, swap.state files, and
large block sizes can add up. Also, most Unix systems save about 5% of the disk
space for root only (so the fs will hit "100%" before all blocks are used).

That, along with reducing fragmentation, is why the FAQ recommends setting the
cache_dir to about 80% of the filesystem capacity.

        -- Brian

On Tue, 29 Aug 2000, Hoomaan Naimi wrote:
>
> Hi all,
> I have configured 3 separate cache directories by means of these tags at squid.config file:
>
> cache_swap_low 95
> cache_swap_high 95
>
> cache_dir ufs /var3/squid 17500 32 256
> cache_dir ufs /var/squid 16384 32 256
> cache_dir ufs /var2/squid 17000 32 256
>
> and this is the report of filesystem disk space usage ( df ) :
>
> Filesystem 1k-blocks Used Available Use% Mounted on
> /dev/hda1 561292 394860 137920 74% /
> /dev/sda3 576504 354380 192840 65% /log
> /dev/sda5 16387643 13470264 2064132 87% /var
> /dev/hda5 17196092 14058808 2263752 86% /var2
> /dev/sdb5 17646672 2450936 14299316 15% /var3
>
> but cache_swap_low and cache_swap_high don't function correctly at /var2. /var2 reaches to its 100% of capacity and after that squid doesn't function correctly and issues these messages continually:
>
> 2000/08/29 00:01:59| diskHandleWrite: FD 8: disk write error: (28) No space left
> on device
> FATAL: Write failure -- check your disk space and cache.log
>
>
> Does anybody knows what is wrong with my configuration?
> Thanks in advance
>
> Best Regards
> Hoomaan Naimi
> Afranet Network Administrator

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