SIP Outbound removed from CE8.x software; how to handle SIP registrar failover?

[Hendrik Schraa]

Hi,

As of CE8.0, Cisco removed the SIP Outbound feature which allowed for SIP systems to register on two SIP registrars/proxies at the same time for fail over purposes (2 clustered VCS systems for example)

Although configuration of multiple SIP proxies is possible, there is no fail over occurring when the first SIP proxy goes down.

Question being of course: how to resolve this?

Current circumstances would not allow for a fail over in case of maintenance or error and i could not find a replacement configuration parameter in CE8.x, any ideas?

Thank you!

[Patrick Sparkman]

I checked an SX80 here in the office and assigned multiple SIP Proxies, however it only showed Proxy 1 as only address being used or even recognized.  Probably the only short term solution is to use DNS for the failover, use a single FQDN address that has several SIP registrar IPs assigned to it.

Regarding the design of CE software and the codec's current behavior in the case of a failover, you should probably open a case with TAC, they can get you the answers you need.

[Justin Ferello]

Hendrik,

Did you ever get an answer to this?

Thanks,
Justin

[Hendrik Schraa]

Hi Justin,

No, unfortunately I've never received an answer to this. I guess if you look at it from a Cisco perspective they will suggest/recommend to switch to UCM based services which handles its own clustering/fail-over of course. But no solution for a VCS based approach in this.

As per the suggestion below, opening up a TAC case might still get clarification, but certainly not an immediate solution.

[Martin Koch]

Hi Justin / Hendrik

Sip outbound is not really failover as its preventing the fail by simultaneous registrations.

I did not check with CE if its in there, in doubt file a feature request, I would go +1 on it.

Regards failover, it should behave the same as old TC endpoints, simply register to 

a hostname as the proxy which points to two or more SRV records of your VCS cluster, 

in case of failure of the current one it would move on to an other one.

For this to happen the old registration needs to expire first, so you want to check the 

SIP Standard registration refresh maximum, the default on the VCS should be 60

which insures if you have a cluster the time it unreachable on inbound calls is small.