cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2956
Views
10
Helpful
13
Replies

Expressway MRA Jabber disconnects every ~45 seconds

sayagi1
Level 1
Level 1

Recently upgraded CUCM cluster to 14SU2 from 11.5 & Expressways from x14.0.3 to x14.0.7.

Now, Jabber devices (Android, iPhone, & Windows) from the outside disconnect & reconnect every 45 seconds or so regardless of it it's on a call or not.

This only affects external devices, not if the device is internal.

Anything off the top of your heads? Log from expressway C when a call was dropped (Im assuming dropped because of the reconnect every ~45 seconds) showed error 422, session timer too small. This happens even for calls within our system such as an external jabber device to an internal desk phone.

No session timer values were changed on our side during the upgrades but this issue appeared afterwards.

13 Replies 13

There is a known defect in CM 14 where you need to add all your C nodes with full FQDN in CM under Expressway C. Easiest is to copy the nodes that are automatically added via AXL in CM and then modify the name of the node to add the domain to form the FQDN.

See this post for additional detailed information on this. https://community.cisco.com/t5/collaboration-applications/cscvz27796-mra-registration-flaps/td-p/4541580



Response Signature


Risat
Level 3
Level 3

Dear, Can you share how you fixed this issue ?

It's all there in my response from 2022-06-30



Response Signature


I saw your reply, just wanted to confirm whether the issue is the same as you mentioned

Infact our issue is resolved now, was caused due to the SIP inspection enabled on the Fortinet firewall, we disabled the same and calls started working now.

Do you know what would be the reason that SIP inspection caused this issue ? SIP inspection would dynamically assign a port for the RTP traffic right, so once the port is set the RTP should be following through the same port as long as the call not disconnected from either side ?

No idea. I don’t know anything about Fortinet.



Response Signature


jari.suosilta
Level 1
Level 1

Did You remember refresh UC servers info in EXP-C?:
We forgot and MRA external connections did start flapping
to go EXP-C
-Configuration
--Unified communications:
---/Unified CM servers /IM and Presencer server nodes /Unity
check box servers box  from  from top left and press REFRESH (note, will drop registration, but hey, they are flapping anyway...)
Do for all configured cucm/cup/unity.
Sometimes really easy to detect if shows old UC Sw version....
Then version info should update.
-------
Mobile and Remote Access Through Cisco Expressway Deployment
Guide (X14.0.2) page 103
Why You Need to Refresh the Discovered Nodes?
This configuration information is static
Changing cluster (such as adding or removing a node)
The information that the Expressway-C reads from the Unified Communications node is different for each
node type/role. These are examples of UC configuration that you can expect to require a refresh from the
Expressway. The list is not exhaustive. If you suspect that a configuration change on a node is affecting MRA
services, you should refresh those nodes to eliminate one known source of potential problems.Check the Unified Communications Services Status
• Changing cluster (such as adding or removing a node)
• Changing security parameters (such as enabling Mixed Mode)
• Changing connection sockets (such as SIP port configuration)
• Changing TFTP server configuration
• Upgrading node software  <<<<<<<<<<<<<<<<< my case

 

 

MiB
Level 1
Level 1

We upgrade from 12.5 to 15 and hit the same bug, solve with using FQDN

MxHadi
Level 1
Level 1

also i suggest update your expressways to version 14.0.11 (which is last sub version in 14.0.x release..) (whiteout losing your license option keys)
maybe this helps to fix if a bug existed..

-- if this helped, please rate by click (Accept as solution) or (Helpful) --

MiB
Level 1
Level 1

I update ucm from 12.5 to 15,exwy was 14.3.4 - hit this bug , upgrade exwy to 15.0.2 - still hit this bug - change hostname to fqdn - works fine

 

It s not related to exwy

It is related to Expw in the sense of how it creates the records for the C nodes in CM via AXL.



Response Signature


Expressway V15.0.1 still creates only the Hostname on CUCM.

Did Cisco plan to fix this on Exwy or on CUCM ?

Not that I’ve seen. One would hope that they do.



Response Signature


MxHadi
Level 1
Level 1

 

New update: (maybe this help)

last day, I was on a project and got an error as this:
external jabber calls on the expressway goes on Preservation mode after 40 seconds, then its disconnected.

I did these two steps and the problem was solved:
1- changed the expressway core hostname to full FQDN in cucm >> device >> expressway-c
2- set DNS + domain name to CUCM + CUP (ssh to them >> set network dns.. + set network domain..) (I forgot to do this..)

-- if this helped, please rate by click (Accept as solution) or (Helpful) --