Hello - we’re having an intermittent issue in that our Asterisk box works fine for hours/days at a time, but suddenly the zap channels go down (busy when dialing in), and the system processes show ‘asterisk’ at 50% CPU usage and ‘events/0’ at 50% cpu usage. If asterisk is restarted at this point, everything goes back to normal, but if one attempts to make an outbound call, the entire server locks up and must be cold restarted.
We’re unable to find any common factor amongst all the times that this is happened (about 5-6), but it does appear that Zap channels are not releasing before it happens. It might be a result of it, or the cause of it, it’s hard to tell. We just notice in the logs that calls start going out/coming in on 2/1 and 3/1 instead of 1/1, when 1/1 should have freed up long before.
This is a new installation that’s only been up for a week and has always had this issue. It’s on a 2.6.26-5 kernel from a Debian kernel image, asterisk version 1.4.21.2 from the Debian apt package, zaptel 1.4.12.1, and wanpipe-3.3.14. Our Zap channels go over a 23 channel PRI line.
Does this sound like an Asterisk issue or a driver issue? Has anyone seen anything like this before? It appears something is triggering off a fast loop or something to eat up so much CPU, but I have no idea where to proceed. Thanks a lot for your time.