[SR-Users] Kamailio Active / Active HA.

[Ben Langfeld]

On 27 December 2014 at 23:44, Carsten Bock <carsten at ng-voice.com> wrote:

> Hi Mahmoud,
>
> Thy typical way, to build a Active/Active setup is by using DNS-SRV
> records. The two servers get different IPs and you announce both IPs
> using DNS to the devices. Most common User-Agents support this
> nowadays, so that's kind of easy. It is specified in RFC3263:
> "Locating SIP-Servers" (http://tools.ietf.org/html/rfc3263).
>
> However, when using DNS-SRV records, dialogs and transactions are
> still associated with a single server, even though the User-Agent has
> an Alternative, in case one server fails. If you want to increase the
> HA for this, you can advertise the Domain name instead of an IP
> (causes more DNS-Lookups).
>

Just a word of caution here, there are several SIP stacks which don't
handle SRV lookups correctly. I don't mean to discourage you from this, but
just to be aware of it. For example, Mobicents only understands SRV record
sets which return domain names. If you return IPs, Mobicents will go ahead
and look them up as A records and fail.


> You can share the dialog state etween two nodes by using the dialog
> module in db-only mode (db_mode = 1, see
> http://kamailio.org/docs/modules/devel/modules/dialog.html#idp15368320).
> You cannot share transactions between two nodes, only dialogs.
>
> In addition, you can work with virtual IPs with Heartbeat/Pacemaker.
>
> At a big operator, we've used a combination of the two: Active/Passive
> for one site and DNS-SRV to distribute it among different sites.
>
> Kind regards,
> Carsten
>
> 2014-12-26 19:07 GMT+01:00 Mahmoud Ramadan Ali <
> cisco.and.more.blog at gmail.com>:
> > Hi Dears,
> > I have successfully configured Kamailio HA using Heartbeat and Pacemaker
> so
> > if one of the two servers should go down the other server will own the
> > virtual IP address and take over.
> > But i have two questions:
> >
> > 1.This model is considered to be Active / Passive redundancy so one
> server
> > will process the SIP signaling until it goes down and i'm wondering is
> there
> > is any way two achieve Active / Active redundancy and if so how the
> > signaling will be handled on the servers so they can be aware of the
> > transactions and dialogs traversing the servers ? does the signalling
> will
> > be replicated or synchronized between the servers or what ?!
> >
> > 2.What about the servers DB and how they should be designed in a cluster
> > mode ? i want to replicate all the DBs of the server to get consistent
> user
> > registration using the subscriber table for example.
> >
> > Thanks in advance and Best regards.
> >
> > _______________________________________________
> > SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
> > sr-users at lists.sip-router.org
> > http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
> >
>
>
>
> --
> Carsten Bock
> CEO (Geschäftsführer)
>
> ng-voice GmbH
> Schomburgstr. 80
> D-22767 Hamburg / Germany
>
> http://www.ng-voice.com
> mailto:carsten at ng-voice.com
>
> Office +49 40 5247593-0
> Fax +49 40 5247593-99
>
> Sitz der Gesellschaft: Hamburg
> Registergericht: Amtsgericht Hamburg, HRB 120189
> Geschäftsführer: Carsten Bock
> Ust-ID: DE279344284
>
> Hier finden Sie unsere handelsrechtlichen Pflichtangaben:
> http://www.ng-voice.com/imprint/
>
> _______________________________________________
> SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
> sr-users at lists.sip-router.org
> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sip-router.org/pipermail/sr-users/attachments/20141229/d2783a9e/attachment.html>