We have a set of new 8841 phones that re-register when going off-hook/on-hook. I've got a TAC case open, however we have not found a cause. Thought I'd post here to see if anyone as seen this behavior from these phones. CM is version 10.5 running on ESXi host. We have other 7962 phones on the cluster that have been working fine for years. Below is a breakdown of the troubleshooting we can tried so far. Any help would be appr
The attached word document has a video to show the behavior.
1. Moved the phone to a different switch, no effect.
2. Attached direct into the switch, bypassing the house network cabling. no effect.
3. Upgraded the phone firmware per Cisco TAC recommendation. no effect.
4. Issue occurs when the phone is registered with the publisher or subscriber.
5. Removed most of the typical references in the phone config, no effect.
6. Upgraded the 2960X network switch software to version 15.0-2a.EX5 per TAC request
7. Disabled jumbo frames on all network switches.
8. Move 8841 to a different CM cluster, problem does not occur.
9. Move known working phone to this call manager, the problem occurs.
Solved! Go to Solution.
You basically have covered everything as far as issue isolation and configuration checks are concerned. I have seen a similar issue once when the Transport Type under Device Security Profile applied on the phone was set to use TCP. The issue was resolved by setting it to UDP. You may try it on a test phone and see if it behaves any differently.
- Do rate helpful posts -
Manish's suspecting the issue is very much acceptable.
Change the protocol from TCP to UDP with his below mentioned changes and test.
Take Packet captures and check for TCP re-transmissions - just look for black packets
Also test by connecting the phone to the switch nearest to the CM. This will help you isolate if this is a phone issue or network related issue and you can troubleshoot accordingly.
(Rate the helpful posts)
I have been having the exact same issue as you are with the 8800 series of phones.
We have narrowed the issue down to a QOS issue with the command service-policy input.
How does the QOS look on your switch?
I found that some of the switch ports had the Cisco soft phone auto qos template applied to the port. This template was apparently dropping SIP packets. Changed it to template to Cisco phone resolved the issue.
I found the exact same thing, in my tests if you take the phone on and off hook you can use up around 200kbps of bandwidth. This is due to the call history and all of the subscribe messages that are sent out to allow for the presence indicators on the phone. If you clear out the call history the bandwidth utilization goes down to what the other Cisco phones produce.
The difference between the softphone policy and the cisco ip phone qos is that the softphone policy policer drops excess traffic where the ip phone policy policer just marks that extra traffic down.