[OpenSIPS-Users] hash and dispatcher on failover
Bogdan-Andrei Iancu
bogdan at opensips.org
Fri Nov 25 16:06:53 CET 2011
Hi Matt,
On 11/25/2011 03:21 PM, Matt Hamilton wrote:
> Thanks a lot Bogdan.
>
> Is it possible to use another ds_select_dst in the failure_route to
> redistribute failed node's calls to remaining Active nodes? Even if I
> can do that, it has to be in the probing_mode=0, am I right? I assume
> I can't "catch" the failure and manually adjust the destination in
> probing_mode=1.
no, you cannot do that as ds_select_xxx() is all the time done over the
original set (like all destinations are active). So you cannot do a
dispatching only over the active ones - dispatcher will compute the
hash, it will identify the corresponding destination and, if failed, it
will shift to the next one.
Regards,
Bogdan
>
> If I can use another ds_select_dst, can it be algorithm 7 with a
> different PV in the failure_route?
>
> Thanks,
> Matt
>
>
>
> ------------------------------------------------------------------------
> Date: Fri, 25 Nov 2011 13:17:08 +0200
> From: bogdan at opensips.org
> To: users at lists.opensips.org
> CC: mistral9999 at hotmail.com
> Subject: Re: [OpenSIPS-Users] hash and dispatcher on failover
>
> Hi Matt,
>
> If Asterisk 4 is down, the hash will not be recalculated, but the
> calls that should go to Asterisk 4 will simply be pushed to the next
> available entry in the dispatcher list.
>
> Regards,
> Bogdan
>
> On 11/25/2011 07:58 AM, Matt Hamilton wrote:
>
> I'm new to dispatcher - wanted to clarify couple of things...
>
>
>
> I'm using algorithm 7 (hash over the content of PVs string) to
> pick a destination.
>
>
> modparam("dispatcher", "hash_pvar", "$pvar")
> ds_select_dst("1", "7");
>
>
> Let's say the probing_mode is 1, and based on the value of $pvar
> (A, B, C, etc.), incoming requests are dispatched in the following
> fashion:
>
>
> A -> Asterisk1
> B -> Asterisk1
> C -> Asterisk2
> D -> Asterisk3
> E -> Asterisk4
> F -> Asterisk4
>
>
> What happens to E, F if Asterisk4 fails? Are they distributed
> among the remaining servers or do they all go to the same server?
>
> Do A, B, C, D still go to the same servers? In other words, is the
> hash recalculated?
>
>
> Can hash_pvar be changed; e.g. use $si for route[1], use $ci for
> failure_route[1]?
>
>
> Thanks,
> Matt
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org <mailto:Users at lists.opensips.org>
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
>
>
> --
> Bogdan-Andrei Iancu
> OpenSIPS Founder and Developer
> OpenSIPS solutions and "know-how"
--
Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
OpenSIPS solutions and "know-how"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20111125/78493197/attachment-0001.htm>
More information about the Users
mailing list