Problems after Upgrade from UCM 8.6.2.20000-2 to 8.6.2.23900-10

[dsobrinho]

Hi guys,

After apply an upgrade of version, we have faced a problem in VMware, the UCM was with high CPU and Memory close to 100%. We have already restart the server before but after some days the utilization backs to high cpu and memory usage.

The tom cat delay is almost 20 seconds to open the web page from CUCM Administrator.

Does anybody know something like that? I have already checked the forum about that, but nothing conclusive.

Following the version

VMware Installation:

        1 vCPU: Intel(R) Xeon(R) CPU           E5430  @ 2.66GHz

        Disk 1: 80GB

        4096 Mbytes RAM

Host Name    : CCCCCCCCCCC

Date         : Fri Nov 8, 2013 14:57:34

Time Zone    : Brasilia Summer Time (America/Sao_Paulo)

Locale       : en_US.UTF-8

Product Ver  : 8.6.2.23900-10

OS Ver       : 5.0.0.0-2

CPU Idle:   00.00%  System:   22.33%    User:   24.27%

IOWAIT:   53.40%     IRQ:   00.00%    Soft:   00.00%   Intr/sec: 593.20

     Memory Total:    4016964K

     Free:                164632K

     Used:               3852332K

     Cached:            1206948K

     Shared:             0K

     Buffers:             103220K

                         Total                     Free            Used

Disk/active         14643120K        2575964K       11918392K (83%)

Disk/inactive      14410496K        2359716K       11902016K (84%)

Disk/logging       50989088K        3425568K       44973344K (93%)

Best regards,

Daniel Sobrinho

[Manish Gogna]

Hi,

You need to run some basic checks on the affected cucm servers.

Please gather the following outputs from the CLI to look for any core dumps, dns, tomcat etc related issues.

utils diagnose test

utils core active list

Is the high CPU being reported during some specific times of the day or is it random? If its at specific intervals then check whether its coinciding with MWI sync, DRS backup, CDR load etc, else RisDC PErfmon logs are required to check the offending processes when the issue is reported.

HTH

Manish

[barry · Accepted solution]

Hi

I've seen something similar when my DRS SFTP server wasn't running and the backups were failing. A restart of the Tomcat service seemed to fix it, rather than a reboot.

Edit: Bizarre. Just seen this again on an 8.6(2) node this morning. DRF Master service was taking high CPU. Restarted DRF Master / DRF Local on the Publisher node and all went back to normal.

Barry Hesk
Intrinsic Network Solutions

[dsobrinho]

Hi Guys,

Problem solved with full restart of server.

I have tried restart the TOM CAT and DRF but did not work. Only after full restart of the VM Proccesss of CPU backs to 20%.

------------------------------

Now, we have new problem. After upgrade UCM 8.6.2.20000-2  to 8.6.2.23900-10, we noticed that the access for user interface of cucm (ccmuser) is not allowing connect.  The integration between Jabber (Webex) and CUCM is not work more and report us problem with Login and password wrong.

Does anybody knows something about that?

Best regards,

Daniel Sobrinho

[Brian Houston]

Daniel

Did you find a resolution to your new problem with ccmuser? I have a customer who has the following issue:

Since we upgraded the CUCM from 8.6.2.21900-5 To 8.6.2.23900-10. Operators/1st Line support users cannot "reset" or "apply config" after saving changes to extension details. A popup comes back saying "user not authorized"

Regards

Brian Houston

[Manish Gogna]

Hi Brian,

Looks like the following bug

https://tools.cisco.com/bugsearch/bug/CSCug29903/?reffering_site=dumpcr

Description

Symptom:

Assigning custom roles to end/application user does not allow to use the Reset/Apply buttons

Hitting save, shows that the update was successful

Conditions:

CUCM 8.6.2.23068-1 or higher

User created Roles

Standard (Built-in) roles should work fine.

Workaround:

Add read/update privileges for the "Voice Mail Pilot web pages" resource to the custom role

HTH

Manish