[Freeswitch-users] Live Upgrade Techniques

[Brian West]

On Jun 11, 2009, at 10:35 PM, John Dalgliesh wrote:

>
> Hi,
>
> On Thu, 11 Jun 2009 at 16:33 -0400, Michael Giagnocavo wrote:
>>
>> Well, if you're running multiple machines, waiting for it to  
>> drainstop
>> isn't that big of a deal unless you're in some sort of hurry, right?
>> Give it an hour or so to drainstop, then kill 'em.
>
> Yes that's exactly what I'm trying to do. The problem is some people  
> will
> only try one IP address.

Clients that don't properly implement SRV/NAPTR and fail over need to  
be smacked.  :)  (not customers but software that fails to do that)


>
>> Would it not be simpler to try to do something with re-invites or  
>> REFER,
>> assuming your endpoints support it?
>
> That was actually plan A. I already added a property in sip_profile  
> called
> failover_redirect, which specifies another server to try if FS can't
> allocate any more sessions (e.g. too busy, paused, shutdown asap,  
> etc.),
> by sending back a SIP 302 Moved Temporarily response, instead of 503  
> Max
> Calls In Progress.

You can't send a 302 to a call thats already established.

>
> Turns out not all my endpoints support it :(

AKA broken endpoints.  :)

>
> I considered REFER too but there seems to be even less support for  
> that.

ACK really?  thats sad!

>
> If I can't get the socket-sharing upgrade working then I will fall  
> back to
> this - and peers which don't support the 302 response (or more likely,
> don't authorise it) will just get no service during the upgrade.
>
>> -Michael
>

Brian West
brian at freeswitch.org

-- Meet us at ClueCon!  http://www.cluecon.com




-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20090611/04f3ccda/attachment.html