[OpenSIPS-Users] Second Contact when sending 302 Redirect using async avp_db_query

[Jon Abrams]

OpenSIPs is using the contents of the RURI for the contact header in the
response.

I seem to recall this working in 2.2 as well:

$ru = "";
append_to_reply( "<sip:" + $rU + "@" + $si + ":" + $sp + ";npdi+">");

If you set the package and/or shared memory larger you can delay the
fragmentation. In my scenario we were using the userblacklist feature and
updating several large lists multiple times an hour. This was triggering
the problem eventually.
You may never run into it with the caching depending on the volume and
allocation patterns. But if you start having response time spikes after a
long uptime, check the logs for memory related messages.

- Jon Abrams

On Fri, Mar 6, 2020 at 5:57 PM Calvin Ellison <calvin.ellison at voxox.com>
wrote:

>
> On Fri, Mar 6, 2020 at 2:56 PM Jon Abrams <ffshoh at gmail.com> wrote:
>
>> Async does indeed trigger the transaction creation. You may need to
>> update the $ru variable and not use the append_to_reply. Or at least that's
>> how I solved this in the past.
>>
>
> What's the connection between the request's URI and a reply's Contact? I
> saw $ru mentioned regarding a similar issue back in 2009 but didn't
> understand how it solves the problem.
>
>
>> I would caution somewhat about using the caching (and possibly the
>> async), as memory fragmentation may become a problem after sustained usage.
>>
>
> Is there any guidance available for how to deal with this? Is this
> addressed via opensips things like hash size or linux things like shared
> memory configuration?
> _______________________________________________
> 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/20200306/498931b0/attachment-0001.html>