Guys,
I've been having a huge problem with deadlocks when using queues with
chan_agent. Testing worked fine, but when I moved it into production, I
started having the system lock up, with openpbx taking 100% CPU.
This is with current OpenPBX pre-0.2 branch on Xen with Linux 2.6.16,
running on a DL/360. We've had very good results with OpenPBX and Xen up
till now.
There is no crash, OpenPBX just runs away with the CPU (which is a pain
when you're using realtime priority). It seems to do that for a few
minutes, then recover, and work perfectly for a while. The only thing
that seems to work while this is going on is ping.
It was doing this with no load on the system, no calls up, and no agents
logged in. We hadn't seen it before trying to use chan_agent, and there
are numerous reports of deadlocks with that code in *.
So I'm currently trying to find a workaround for the desired behaviour
(using AddQueueMember will keep calling agents while they're already on
the phone), but thought it would be good to make people aware of this.
---Nathan
_______________________________________________
Openpbx-dev mailing list
[email protected]
http://lists.openpbx.org/mailman/listinfo/openpbx-dev
