We are having to reset H323 gateways at remote sites every 2-3 weeks it seems, not all just a few. The sites complain that they are unable to make outbound calls so we reset the gateway in call manager and the issue is resolved. Has anyone else experienced this and found a final resolution? We are running CCM 5.1.3
I should have read a little better. I noticed there are quite a few (90+) H323 bugs for your version of CM.
This could be your issue:
It might be worth moving to the newest service release.
I doubt that this is a gateway issue.
H323 gateways do not have any type of persisting connection or registration to CUCM. If there is an ongoing call there may be a TCP connection up on 1720, but that is it.
I think there is a higher likelihood that the issue is in CUCM with regards to the route pattern/list and hunting algorithms.
Cisco TAC responded with a similiar response. h225d component is failing for the gateways that have issues. Reset creates new h225 process in CUCM. Thank you for the response.
We have the same problem at our customer. Users at certain sites can't call out via H.323 gateway located there, and this occurs randomly, probably anywhere, at any site.
What was the solution for your case?
Did you upgrade to the latest 5.1.x, or upgrade further to 6.x?
Thank you in advance!
No resolution as of yet. TAC is waiting for the next occurence so we can debug, which is kind of hard because our sites can't receive customer calls.
Did you have a resolution to this issue? We had two instances in last week with same issue where outbound calls did not work while inbound calls worked. Believe the call setup could not make it to the gateway as there was no Q931 logs. Reset of the H323 gateway on CUCM 8.0 fixed the problem both the times. Please answer with steps that resolved this issue for you.
we realized that when we moved to the new office, we renamed the old gateways in the cucm to new gateways (rename as in new ip). that seem to have caused some confusion in h225d instance randomly . so i deleted the gateways in cucm and added them back followed by the cucm cluster reload. havent had any issues since then.