Varnish 2.06 running away with memory

[Laurence Rowe]

Be aware that if you use varies, you must purge all potential
combinations (for example, variations on Accept-Encoding) individually
when using `set obj.ttl = 0s`. I believe Varnish 2.1 now has a thread
to process the purge list and prevent it growing out of hand.

Laurence

On 18 May 2010 15:41, Sam Crawford <samcrawford at gmail.com> wrote:
> Scott,
>
> Thanks for that, I'll experiment with that and see how we fare!
>
> Regards,
>
> Sam
>
>
> On 18 May 2010 15:19, Scott Wilson <scott at idealist.org> wrote:
>> Hi Sam,
>>
>> On 18 May 2010 10:47, Sam Crawford <samcrawford at gmail.com> wrote:
>>> Morning all,
>>>
>>> A bit of a newbie question I fear... We've been successfully running
>>> Varnish 2.04 for over a year now on our intranet, and have found in
>>> the past two days that our instances in one particular region are
>>> eating up a lot of system memory. We've upgraded to 2.06 a couple of
>>> days ago, and are still experiencing the same issue.
>>>
>>
>>> n_purge                246879          .   N total active purges
>>> n_purge_add            252285         5.61 N new purges added
>>> n_purge_retire           5406         0.12 N old purges deleted
>>> n_purge_obj_test       245731         5.46 N objects tested
>>> n_purge_re_test     214813964      4773.11 N regexps tested against
>>> n_purge_dups             3227         0.07 N duplicate purges removed
>>> hcb_nolock                  0         0.00 HCB Lookups without lock
>>> hcb_lock                    0         0.00 HCB Lookups with lock
>>> hcb_insert                  0         0.00 HCB Inserts
>>> esi_parse                   0         0.00 Objects ESI parsed (unlock)
>>> esi_errors                  0         0.00 ESI parse errors (unlock)
>>>
>>
>> It looks to me like you're generating an awful lot of purge list
>> entries.  We struggled for a long time with this problem before
>> abandoning purge in favour of setting object TTLs to 0.  See this FAQ
>> entry:
>>
>> http://varnish-cache.org/wiki/VCLExamplePurging
>>
>> The problem we were having was that we were using purge to implement
>> force-refresh functionality, but there are so many badly behaved
>> crawlers and other clients out there that send always send the NoCache
>> header that the purge list would swamp all available resources.
>>
>> Hope that helps!  cheers,
>> scott
>>
>
> _______________________________________________
> varnish-misc mailing list
> varnish-misc at varnish-cache.org
> http://lists.varnish-cache.org/mailman/listinfo/varnish-misc
>