cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
308
Views
0
Helpful
2
Replies

AS5300 reload to free pending channels in trunk group

javiero_2
Level 1
Level 1

I have two AS5300's with 4 T1-CAS controllers in a trunk group. These AS5300's are used as outbound VoIP gateways for PSTN access. Both of these AS5300 have a problem that occurs about three times a day and they need to be reloaded. There is always approximately 80-85 calls up on each gateway on the trunk group. When the gateways are running normal and I do a 'show trunk group 1' I get the following summary:

Total calls for trunk group: Data = 0, Voice = 72, Modem = 0 Pend = 15, Free = 9

Over a period of time the number of pending channels increments until there are no free or in use channels. For example:

Total calls for trunk group: Data = 0, Voice = 11, Modem = 0 Pend = 85, Free = 0

These channels are actually available but they are somehow staying stuck in pending after a period of time.

Both of the gateways have had numerous versions of code on them from 12.1 to the current one on both gateways 12.3(5b). I have seen similar like bugs on the Bug toolkit but do not think any of them really apply to this issue since the workarounds and fixed code has not resolved these issues.

Maybe there is another way for me to utilize all four T1's by a single pots dial-peer and not using a trunk group as a target. It appears that this issue is caused by the trunk group resource manager but I cannot get a definite answer. The constant reloading around the clock is getting tiresome. I appreciate anyone's input who may have had a similar issue. Thanks

Javier

2 Replies 2

ivillegas
Level 6
Level 6

The pending state is cause by an unexpected interuption of the call between the gateway receiving the Voip call setup and sending a isdn setup. The gateway reserves a circuit on the E1 by putting it in pending state, but then does not proceed with the call, for some external reason, either due to the remote gateway ending the call, or perhaps a glare situation. Either way the statemachine doesnt release the pending state, if the call fails at this precise time. I think this a bug.

Thanks for the info.

Javier