Dear All,

I get wccp working after shrink the cache size of rock store from
170GB down to 70GB. My server memory is 32GB.

# Rockstore filesytem
workers 4
cpu_affinity_map process_numbers=1,2,3,4 cores=2,4,6,8
# Rock Store: SMP Awared
cache_dir  rock /cache1         70000 max-size=31000 max-swap-rate=300
swap-timeout=300
cache_dir  rock /cache2         70000 max-size=31000 max-swap-rate=300
swap-timeout=300
cache_dir  rock /cache3         70000 max-size=31000 max-swap-rate=300
swap-timeout=300

# AUFS file system
if ${process_number}=4
cache_dir  aufs /cache4/squid/${process_number}         170000 32 256
min-size=31001 max-size=200000000
cache_dir  aufs /cache5/squid/${process_number}         170000 32 256
min-size=31001 max-size=200000000
cache_dir  aufs /cache6/squid/${process_number}         170000 32 256
min-size=31001 max-size=200000000
cache_dir  aufs /cache7/squid/${process_number}         170000 32 256
min-size=31001 max-size=200000000
cache_dir  aufs /cache8/squid/${process_number}         170000 32 256
min-size=31001 max-size=200000000
endif

---
Regards,
Vantha

On Wed, Mar 20, 2013 at 9:44 PM, Alex Rousskov
<rouss...@measurement-factory.com> wrote:
> On 03/20/2013 12:58 AM, Amos Jeffries wrote:
>
>> I am suspecting the problem is related to the WCCP default of waiting
>> until all caches are loaded before starting to advertise HERE_I_AM.
>> Scanning 1.4 TB of disk is going to take a while. Sokvantha YOUK was
>> waiting _only_ about ten minutes for WCCP packets.
>
> I see. This behavior sounds like a Squid bug to me: By default, Squid
> should advertise its presence when it is ready to service requests, not
> when it is done loading the cache index. In [rare] situations where the
> cache index is required, there is a "foreground rebuild" option (IIRC)
> that should be taken into account when fixing this.
>
>
>> With the "fixed" config the AUFS disk cache is isolated into a worker
>> separate from the rock stores. This introduces a few new factors:
> <snip>
>
> Agreed. However, since that config is wrong, we should probably ignore
> its side-effects especially if we can now explain them.
>
>
> Thank you,
>
> Alex.
>



-- 
----
Regards,
Vantha

Reply via email to