Re: [squid-users] wccp service lost issue

From: Amos Jeffries <squid3_at_treenet.co.nz>
Date: Fri, 21 Aug 2009 19:57:10 +1200

Vivek wrote:
> Amos,
>
> I didn't see any errors related to squid restart/crash in cache.log. Is
> there any way to debug this ?
>

Other than dumping the WCCP info display on the router I have no idea.

As Adrian said, its probably just a temporary issue with a long index
rebuild causing Squid to ignore one of the WCCP announcements.

If so the loss time will remain the same no matter how frequently or
infrequently you rotate. Thus less often will be less outage time.

>
> -----Original Message-----
>
>
> Tom Penndorf wrote:
>
>> Hello,
>
>> Vivek schrieb:
>
>>> Amos,
>
>>> I am rotating the logs using " squid -k rotate ". In the crontab,
>
>>> 0 */1 * * * /usr/local/squid/sbin/squid -k rotate
>
>> I think, this will cause squid to do an reload, so it will shortly
> stop
>
>> the wccp-service. Perhaps, you should rotate the files only once a
> day.
>
> "-k rotate" was the right way to do it. Only reloads the helpers and
>
> rotates the logs.
>
> WCCP and other public-facing services should not be seeing any loss of
>
> service at all on rotate. Worst case is a short[1] lag while the store
>
> index journal gets rebuilt and dumped to disk.
>
> [1: for various definitions of 'short'.]
>
> Is there any sign in cache.log of Squid crashing at or around those times?
>
> 2.7 has an auto-restart built in that may be masking it from notice.
>
>
>
>>
>
> Tom
>
>>
>
>>> squid version - squd-2.7 Stable 6
>
>>>
>
>>> IOS version - 12.4(15)T6
>
>>>
>

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE6 or 3.0.STABLE18
   Current Beta Squid 3.1.0.13
Received on Fri Aug 21 2009 - 07:57:19 MDT

This archive was generated by hypermail 2.2.0 : Mon Aug 24 2009 - 12:00:04 MDT