[SR-Users] Interactions of dmq_usrloc and usrloc w/db_mode > 0
arsperger at gmail.com
Mon Sep 26 20:17:53 CEST 2022
by the chance I was playing exactly with the same setup these days, in a
lab everything works just fine. +1 to the initial question.
On Mon 26 Sep 2022 at 18:33, Joel Serrano <joel at textplus.com> wrote:
> +1 in this situation haha, also hoping to get some nice input :D
> On Mon, Sep 26, 2022 at 9:20 AM Matteo Brancaleoni <mbrancaleoni at gmail.com>
>> I asked a similar question here
>> https://lists.kamailio.org/pipermail/sr-users/2022-July/115160.html but
>> no answer yet :)
>> What I see on my side is that it indeed works, the only drawback is that
>> the same contact is getting synced to DB, which causes a duplicate error
>> unless you use the "db_insert_update" option.
>> Done that seems that it works ok (did tested in prod yet), and I had the
>> all the contacts live on all nodes and on db. The only downside that maybe
>> can happen is that the periodic sync may skew a bit the expire time and
>> maybe give a contact some more seconds, but it really depends on timings of
>> the clusters.
>> But I'm really interested in the answer, too :)
>> -- Matteo
>> On Mon, Sep 26, 2022 at 4:07 PM Alex Balashov <abalashov at evaristesys.com>
>>> Are there any known contraindications for replicating contacts using
>>> dmq/dmq_usrloc, and injecting those contacts into a database on one of the
>>> nodes using usrloc with `db_mode` 1 or 2?
>>> Predictably, this is being done to support the use-case of getting
>>> registration status from database. If you think this should be done with
>>> JSONRPC, I am in complete agreement with you, but it’s not up to me. :-)
>>> I am doing this with `db_mode` 2 now, and finding that, for a small
>>> number of AORs, one can find instances where they are consistently stored
>>> in memory but not present in the `location` table. The overall proportion
>>> of these AORs, out of thousands, seems to be quite small. It was initially
>>> somewhat higher, and it went down once I increased usrloc `timer_processes`
>>> and increased the sync interval from 30 to 60 seconds.
>>> Nevertheless, it is still non-zero, and I am getting intermittent
>>> reports. I wonder if there are some prior experiences with this and
>>> anything to watch out, or if `db_mode` 1 might be a superior choice. I
>>> personally cannot see how that would be; it seems to have all the
>>> performance downsides of mode 3. But perhaps if something about it is more
>>> “problem-free” vis-a-vis DMQ, it’s worth a shot?
>>> — Alex
>>> Alex Balashov | Principal | Evariste Systems LLC
>>> Tel: +1-706-510-6800 / +1-800-250-5920 (toll-free)
>>> Web: http://www.evaristesys.com/, http://www.csrpswitch.com/
>>> Kamailio - Users Mailing List - Non Commercial Discussions
>>> * sr-users at lists.kamailio.org
>>> Important: keep the mailing list in the recipients, do not reply only to
>>> the sender!
>>> Edit mailing list options or unsubscribe:
>>> * https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users
>> Kamailio - Users Mailing List - Non Commercial Discussions
>> * sr-users at lists.kamailio.org
>> Important: keep the mailing list in the recipients, do not reply only to
>> the sender!
>> Edit mailing list options or unsubscribe:
>> * https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users
> Kamailio - Users Mailing List - Non Commercial Discussions
> * sr-users at lists.kamailio.org
> Important: keep the mailing list in the recipients, do not reply only to
> the sender!
> Edit mailing list options or unsubscribe:
> * https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users
Sent from Gmail Mobile
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the sr-users