I have just replaced an old server stuck on 1.2 asterisk due to prerequisites with a TE405 quad T1 card. The new machine is running suse with a TE420 quad port T1 card, 1.8.9.1 asterisk and 2.6.0 dahdi. One PRI is connected to our local provider (all of our inbound calls to this building arrive on this circuit), a second is connected to our long distance provider.
Occasionally we’ll place several thousand calls that are maintained to about 40 simultaneous outbound calls for a couple of hours. The call lists tend to be more local calls. This combined with additional inbound calls will tend to keep all 23 channels in use on our local PRI while excess local calls roll over to the long distance PRI. Inbound calls might be rejected when this occurs.
We have used our automated dialer twice since the server was upgraded. On both occasions by the end of the process, the local PRI would not accept inbound calls - even after the automated calls were completed but outbound calling worked normally. Stopping and restarting asterisk or physically unplugging and replugging that T1 restored normal operation.
I was in a hurry to get the system running so didn’t have time to collect data. But if I placed a call the message on the cli (verbose 3) was an ISDN 33 code on the outbound channel - there was no cli on the failed inbound half.
module reload chan_dahdi.so # did not help
dahdi restart # ended asterisk (I think)
the chan_dahdi.conf for this circuit is as simple as it gets…
switchtype=national
signalling=pri_cpe
context=ec
group 1
channel=>1-23
The only real difference between the new & old configuration is I’m trying oslec instead of one of the mec echo cancelers.
It looks like it may be related to issues.asterisk.org/jira/browse/DAHLIN-268