Re: Jabber 14 Android MRA login to IM problem

pgolub · ‎07-28-2021

I have the following setup: CUCM 11.5(1)SU10 IMP 11.5(1)SU10 both clustered,

MRA based on two Exp-C and Exp-E clusters of 14.0.2

Everything worked find for long time unlit users devices started receiving Jabber 14 update.

The only problem I have with Android devices running any release of Jabber 14. After some time (it could be hours or days) Jabber IM service going to DISCONNECTED state ( "Cannot communicate with the server" message on JBC )and stay like this until next login. Phone service keeps working.

We use simple credential authentication (no SSO).

If I install Jabber 12.9.7 on the same Android phone the problem disappears.

Debug of XCP Authentication service attached.

Basically, JBC getting something like:

And it happens several times until JBC gets disconnected. Client can't just get new onetime password for xmpp.

Any suggestions are appreciated

Miguel Castillo (miguecas) · ‎09-14-2021

Hi,

As you mentioned, the problem seems to reside on the server unable to process and generate the ONE-TIME-PASSWORD.

I would suggest checking the CLI outputs of utils imdb_replication status from both IM&P nodes, as it looks there is an issue with the IMDB.

The suggestion is to:

Step 1. Disable the High Availability from the Presence Redundancy Groups under CUCM Administration > System.

Step 2. Stop the Cisco Presence Engine on all the IM&P nodes.

Step 3. Verify all the Datastore Services are running: Cisco Login Datastore, Cisco Route Datastore, Cisco Presence Datastore, Cisco SIP Registration Datastore.

Step 4. Restart the Cisco Config Agent on all the IM&P nodes, each node at a time.

Step 6. Restart the Cisco XCP Authentication Service n all the IM&P nodes, each node at a time.

Step 5. Start the Cisco Presence Engine.

Step 6. Reenable the High Availability

Regards, Miguel Castillo

pgolub · ‎09-16-2021

Hi Miguel,

Thanks for your reply.

As you suggested, I checked IMDB replication from both IM&P nodes. No problems with it.

However, I have done all the above steps and checked IMDB replication again. All datastore replications were passed.

Then I tested Jabber for Android 14.0.3.306226 and found that the problem with ONE-TIME-PASSWORD still exist.

If I use Jabber for Android 12.9.7.305475 then there is no such problem.

Regards, Paul

Valeriy Kuznetsov · ‎09-17-2021

Hello Paul,

Having only XCP Auth Service is not enough to say why it is failing. OTP generated by Client Profile Agent service.

We would need below trace set from IMP for failed auth attempt:

Cisco Tomcat

Cisco Tomcat Security

Client Profile Agent (set to Debug!)

XCP Authentication Service

On top of that since this is MRA login we might need diagnostic logs from Expressways, plus Problem Report from Jabber (with enabled detailed logging).

But for basic troubleshooting I can take a look for IMP trace set mentioned above.

Regards,

Valeriy

pgolub · ‎09-19-2021

Hi Valeriy,

Thanks for your reply.

I have attached the requested logs. Unsuccessful attempt to get OTP for jid pgolub was around 4 pm on Sunday.

Regards,

Pavel

Valeriy Kuznetsov · ‎09-20-2021

I checked the logs, cannot see a problem on IMP side.

The thing is that Jabber tried multiple times to connect to IMP over Expressway in between 16:00 and 16:04, those connections were dropped in the middle. After few attempts Jabber client resigned from further connection.

What i can see that android device is switching back and forth between WI-FI or LTE network, getting new ip-address etc, network is not stable.

/// 
2021-09-19 16:00:18,665 INFO  [IMPServices] [updateServerInfo] - *********** IMP Outage::Setting server details to : host is expe02.stec.ru and port is : 5222
2021-09-19 16:00:18,666 DEBUG [IMPServices] [updateServerHealth] - *********** IMP Outage::Setting server health to Reconnecting
2021-09-19 16:00:18,774 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged conn, network changed or network down. err:24
							  
2021-09-19 16:00:30,340 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged current local address:10.36.25.183, new queried local address:10.36.25.183
							  
2021-09-19 16:02:44,190 DEBUG [IMPServices] [setServerHealth] - *********** IMP Outage::Setting server health to Disconnected
2021-09-19 16:02:55,927 INFO  [IMPServices] [updateServerInfo] - *********** IMP Outage::Setting server details to : host is expe02.stec.ru and port is : 5222
2021-09-19 16:02:55,928 DEBUG [IMPServices] [updateServerHealth] - *********** IMP Outage::Setting server health to Connected
							  
///                           
							  
2021-09-19 16:03:06,200 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged conn, network changed or network down. err:24
2021-09-19 16:03:06,220 DEBUG [IMPServices] [setServerHealth] - *********** IMP Outage::Setting server health to Reconnecting
2021-09-19 16:03:15,611 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged current local address:10.207.99.249, new queried local address:10.207.99.249
2021-09-19 16:03:19,079 DEBUG [IMPServices] [updateServerHealth] - *********** IMP Outage::Setting server health to Connected
							  
///                           
							  
2021-09-19 16:03:28,935 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged conn, network changed or network down. err:22
2021-09-19 16:03:28,935 INFO  [csf.jwcpp] [onNetworkChanged] - @XmppSDK: #0, onNetworkChanged current local address:10.207.99.249, new queried local address:10.3.129.111
2021-09-19 16:03:28,938 DEBUG [IMPServices] [updateServerHealth] - *********** IMP Outage::Setting server health to Reconnecting

pgolub · ‎09-28-2021

Can you please have a look at the attached logs again. No IMDB replication errors but JBC got into the same state when IM isn't connected and SIP is connected. It happened at 14:43 on September 28. Here's XCP auth log event:

Thank you in advance.