[OpenSIPS-Users] Memcache issues

Vlad Paiu vladpaiu at opensips.org
Tue Feb 5 12:57:31 CET 2013


Hello,

About the memcached TIMEOUT issue, I have tested the scenario where the 
memcached server doesn't reply, and it seems to be working, in the sense 
that the cache_fetch function returns a negative response code to the 
script, so "if (!cache_fetch())" should work. Could you send me the logs 
from when the memcached gives timeout, along with the relevant snippet 
from your OpenSIPS script ?

For the distributed dialog profiles, I have committed in 1.8,1.9 and 
trunk a fix that solves the prefix issue.

We'll be looking at the ratelimit issue and get back to you.

Best Regards,

Vlad Paiu
OpenSIPS Developer
http://www.opensips-solutions.com


On 02/05/2013 12:32 PM, Brett Nemeroff wrote:
> They are separate issues.
>
> I'm getting regular memcache errors like SYSTEM ERROR or TIMEOUT for 
> just normal script cache_fetchs. They don't happen a whole lot, but 
> when it does, I get like 10-15 of them all at once. And like I said, 
> it doesn't appear to return negative so call processing kinda fails 
> instead of letting me do the proper action when the key is missing.
>
> The errors I'm getting with distributed dialog profiles and 
> ratelimiting are different. I'm using the prefix modparams for dialogs 
> and it appears to be truncating those prefixes oddly. Ratelimiting 
> doesn't appear to have the same prefix issue. Ratelimiting however 
> also appears to be broken because I have a perfectly idle box showing 
> counters in rl_list. I'm not really sure what's going on and it's hard 
> to troubleshoot because ratelimiting inserts a very large integer to 
> the backend and I'm not sure what it represents.
>
>
>
> On Tue, Feb 5, 2013 at 4:27 AM, Bogdan-Andrei Iancu 
> <bogdan at opensips.org <mailto:bogdan at opensips.org>> wrote:
>
>     Brett,
>
>     I understand you are working with Vlad in troubleshooting the
>     memcached issues.
>
>     But if I understand correctly, you are saying that you are also
>     getting errors when using distributed dialog profiles but with
>     couchbase ? If so, what kind of error you get there ?
>
>     Regards,
>
>     Bogdan-Andrei Iancu
>     OpenSIPS Founder and Developer
>     http://www.opensips-solutions.com
>
>
>     On 02/05/2013 11:52 AM, Brett Nemeroff wrote:
>>     Bogdan,
>>     Increasing the -R parameter seems to have made the situation
>>     better, but I am without question regularly running into weird
>>     memcache timeout and system errors. I've tried several different
>>     hardware platforms. Point to local and remote memcache servers.
>>     Different memcache libraries and server versions. I've tried
>>     couchbase instead of memcache and I can't get rid of the errors.
>>     I'm really not sure what could be causing them.
>>
>>     When I get them, the call attempting to process fails, which is
>>     frustrating. I'd expect that the memcache result would return
>>     negative which would trigger a direct db query (which is written
>>     in the script in case the cache object is missing anyway).
>>     However it doesn't do this. It just fails and continues
>>     processing a null return value. Is that behavior expected or a bug?
>>
>>     Thanks!
>>     -Brett
>>
>>
>>     On Thu, Jan 24, 2013 at 6:27 AM, Bogdan-Andrei Iancu
>>     <bogdan at opensips.org <mailto:bogdan at opensips.org>> wrote:
>>
>>         Hi Brett,
>>
>>         We talked about this on IRC - any luck with it ?
>>
>>         Regards,
>>
>>         Bogdan-Andrei Iancu
>>         OpenSIPS Founder and Developer
>>         http://www.opensips-solutions.com
>>
>>
>>         On 01/21/2013 11:31 PM, Brett Nemeroff wrote:
>>>         Hey All,
>>>         Generally I have really good luck with memcache, but for
>>>         some reason, I'm having some issues today.
>>>
>>>
>>>         Once I crank this one box up with load, I start to see these
>>>         errors flood my logs:
>>>
>>>         cachedb_memcached:wrap_memcached_get: Failed to get: SYSTEM
>>>         ERROR
>>>         cachedb_memcached:wrap_memcached_get: Failed to get: A
>>>         TIMEOUT OCCURRED
>>>
>>>         I've tired changing the tread count for memcache. Additional
>>>         processes.. Heck, I've even tried putting memcache on other
>>>         servers and in other colos. The original memcache is on the
>>>         same box and is pointed at localhost (127.0.1.2).
>>>
>>>
>>>         I'm hitting it reasonably hard, but it should keep up. About
>>>         10k ops a sec. I see it run great for like 2 minutes, then
>>>         it hangs for like 4 seconds, then picks back up for another
>>>         30 seconds or so.. goes that like on and off.
>>>
>>>         Any idea of what could be wrong?
>>>
>>>         Thanks,
>>>         Brett
>>>
>>>
>>>         _______________________________________________
>>>         Users mailing list
>>>         Users at lists.opensips.org  <mailto:Users at lists.opensips.org>
>>>         http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>
>>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20130205/c2db8411/attachment.htm>


More information about the Users mailing list