[asterisk-dev] distributed device-state implementation question

[Russell Bryant]

On Thu, 2011-01-13 at 11:45 -0600, Russell Bryant wrote:
> On Thu, 2011-01-13 at 16:47 +0100, Klaus Darilion wrote:
> Excellent question.  The device state cache (which is independent of the
> state distribution method) will remember the last known state of a
> device on a given server.  If that server never comes back up, the state
> on that server will be stuck at busy, which would mess things up.

Actually, I don't think this applies to the SIP method being developed,
since the state collection isn't done inside of Asterisk.  These
comments only apply to the XMPP and AIS methods.

> At least with AIS, we get notified when a node joins or leaves.  We
> could use that as a trigger to clear all cache entries for a given
> server, but we do not do that right now.  I don't know the best way to
> deal with it for other distribution methods.


-- 
Russell Bryant
Digium, Inc.  |  Engineering Manager, Open Source Software
445 Jan Davis Drive NW   -    Huntsville, AL 35806  -  USA
jabber: rbryant at digium.com    -=-    skype: russell-bryant
www.digium.com -=- www.asterisk.org -=- blogs.asterisk.org