Cisco Meeting Server XMPP Authentication Service Failure

[msamadpour]

We are currently on CMS version 2.0.6. Single-Split deployment with a single XMPP server and an odd number of databases (per Cisco best practice)

We are detecting issues in which CMA apps, CMA iOS apps, etc. that require user authentication are failing sporadically. We have narrowed the issue as being the authentication component for XMPP. 

When we log into the webadmin of the XMPP node, we see logs below:

10:54:29:  WARNING : failed to register with authentication component "authp-standalone.fqdn.com" -- request timed out

10:54:29:  INFO : destroying registration to authentication component "authp-standalone.fqdn.com" -- not discovered

11:04:31:  INFO : registration with authentication component "authp-standalone.fqdn.com" succeeded

The service stops on its own and then ultimately restarts itself. The issue is generally cleared up be restarting XMPP manually on the node. 

Has anyone else run into this issue? Is this a bug?

[Ammar Saood]

did you create callbridge component in XMPP ?

webadmin and callbridge are almost same component.

webadmin needs to connect to XMPP.  so you tell the XMPP to which callbridge to connect to.

xmpp callbridge add c_Call

it will generate a secret.  say "xxmxywa323zxxx"

now add this secret in webadmin under Xmpp 

unique call bridge name = c_call

server address =  interface where XMPP is listening  on port 5223

shared secret =xxmxywa323zxxx

you can manually specify the server address with port if callbridge and XMPP are listening on same interface

127.0.0.1:5223  or 192.168.0.50:5223 

both are valid addresses.

HTH 

AMMAR

Please remember to rate responses and to mark your question as answered if appropriate.

[msamadpour]

Hi Ammar - 

Thanks for the suggestions. Please note that the call bridges and XMPP configurations are all set and working. 

The XMPP authentication service appears to simply crash and restart again.

[Ammar Saood]

upgrade to CMS 2.08.

snapshot confirms that XMPP and callbridge connected but not authenticated.

i will suggest you double check the Shared secet +xmpp domian you configured on XMPP component in CLI. it must match with Webadmin settings.

Please remember to rate responses and to mark your question as answered if appropriate.

[msamadpour]

Hi Ammar - this had nothing to do with the configuration. It appeared to be a bug on the 2.0.6 platform which has yet to be confirmed or documented. After upgrading the platform to 2.0.8, the issue has subsided.

[Ammar Saood]

good to know its fixed now.

[msamadpour]

For the record, this appears to be some sort of undocumented bug on the 2.0.6 platform. After upgrading the platform to 2.0.8 (no configuration changes made) - the XMPP authentication service has stabilized.

[George Thomas]

Has this been solid for you? I am having the same issue, running the latest and greatest version.

[msamadpour]

How many nodes in your environment? Number of XMPP servers?

Did you have this issue prior to your recent version?

This has been solid for us since upgrading to version 2.0.8. I have not moved past this version since the platform stabilized?

[George Thomas]

Thanks for the response, its 3 nodes with 3 XMPP servers. I had issues with 2.0.6, 2.0.9 and now on 2.1.2.

[msamadpour]

We have a single XMPP node and ran into the issues on all version prior to 2.0.8

Do you have any opportunity to rollback to v2.0.8 to validate?

If unable, I would highly suggest opening a TAC case and specifying the XMPP authentication failure as being the root cause of failure. 

The issue can be then pushed up to the BU - since this appears to be more than just a single case (my issue from a few months ago)

[George Thomas]

I have opened a case with TAC and getting it escalated. I removed the Callbridges from XMPP (based on another person's recommendation) and re-added it back that seems like to have made it stable (knock on wood).

[Ammar Saood]

i never had such issue.

1. make sure no 2 IP addresses on CMS server are in same subnet it causes routing issues.

2. may be the secret was added with spaces ? 

it happens with me many times, directly copying the something from Putty and pasting into an application, adds spaces at times.copying into notepad first solves my issues.

you can upgrade to 2.1.2. its all stable.

keep us posted with what TAC says.

THANKS

[Anthony_alc]

Hello everyone

I had the same problem, XMPP authentication failure. I test it in 2.0.8 , 2.1.2 and 2.1.3 , all version with same error. 3 CMS with 2 interfaces , CMS01 int a webadmin, webbridge,XMPP int b DDBB callbrige , CMS02 int a webadmin, webbridge,XMPP int b DDBB callbrige , CMS03 int a webadmin, XMPP int b DDBB all in the same subnet.

Then I deployed a new CMS VM server with all services in the same interface, same error, but I noticed something, there was a error when I added the callbridge to the XMPP service.

CMS01> xmpp callbridge add callbridgeCMS01
Success             : false
Info                : Parent IM domain <CMSSERV.stgcollab.com> of domain <callbridgecms01.cmsserv.stgcollab.com> does not exist

So I decided to review all XMPP setting and the domain had some upper case letters

Domain                  : CMSSERV.stgcollab.com

I changed it to

Domain                  : cmsserv.stgcollab.com

and the XMPP registered

In the file attached you can find some settings what I applied in the XMPP leader.

Regards

[George Thomas]

After some back and forth with TAC/BU, it was identified as a bug. Here is bug ID: CSCve09761. We have a ES for it which seems to be working fine, official release is slated to come out this week.