In this article we discuss some issues of the CUCM environment troubleshooting.
This time we will consider the aspects of CUCM availability.
So, your CUCM responds slowly or doesn’t respond at all. Why?
CUCM does not respond to endpoint requests
When the Primary CUCM slows down or even dies IP phones and/or gateways lose the registration. When taking the receiver off the hook the tone appears with delay or does not appear at all.
The most likely reasons for this are:
- The CUCM server hung on the OS level and requires a reboot.
- The Cisco Call Manager service hung. Also, there may be problems with the Cisco TFTP Server service responsible for the phone configuration loading. Check the status of both services (Serviceability > Tools > Control Center> Feature Services). Causes of the problems can be found in System logs .
- High CPU load of the CUCM server. Check the CPU Usage to understand what process should be blamed.
- A memory leak - you need to check the memory usage.
CUCM System logs
CUCM is an appliance based on Linux, and it does not provide a regular access to a full Linux console, therefore System log can only be viewed through RTMT.
We need a Syslog Viewer - the analogue of Windows Event Viewer. If something is wrong with the system it’s one of the first interfaces to look at.
System > Tools > Syslog Viewer
Syslog Viewer allows you to view the messages from the following logs:
- System Logs – everything that concerns hardware and OS
- Application Logs – CUCM logs
- Security Logs – user login attempts etc.
Read more about working with RTMT here: Working with Cisco Unified RTMT
If you have problems with the Cisco Call Manager service, then the following messages can appear.
When the connection with the service is lost:
The Cisco CallManager service terminated unexpectedly. It has done this one time. The following corrective action will be taken in 60000 ms. Restart the service.
Timeout 3000 milliseconds waiting for Cisco CallManager service to connect.
If you have problems with the launch:
The service did not respond to the start or control request in a timely fashion.
Cisco Call Manager service may stop responding because of a system resources (processor or memory) overloading.
Most often it occurs at a high CPU load.
We can monitor the CPU load in two ways:
- CPU monitoring in CLI.
The average CPU load can be seen by executing:
show stats io
show perf query class Processor
CPU load by processes:
show perf query counter Process "% CPU Time"
show process load
- CPU monitoring in RTMT
This can be checked in the general information on the system:
System > System summary
CUCM Administration page is not displayed
If the CUCM admin page (https:///ccmadmin) does not open:
- First try to clear the browser’s cache
- Check the network settings, try to ping the CUCM IP
- If the CUCM is accessed by name, check the DNS, and whether the name resolves correctly
- Make sure the Cisco Tomcat service is running
- Check Firewall and Access Lists settings
- Check the CPU load on the CUCM node
Checking Cisco Tomcat service
Use CLI to check the status of the service:
utils service list
Launch it if necessary:
utils service start Cisco Tomcat
Slow Server response
The slow server response may cause: the delay to receive the dial tone, delays in opening the admin/user web-interfaces, delays in dialing.
The possible reasons are:
- The CUCM server and the switch have different Speed / Duplex.
Check the settings on the server and the switch. Try to set Auto for both.
- High CPU load or Memory leak. Check the CPU usage.
- Also, the wrong Dial plan may cause delays in dialing.
The CUCM network settings may be checked with the CLI command:
show network eth0
You can change the value of the duplex or speed in CUCM by executing:
set network nic eth0
To view the switch network settings execute:
show interface fa 1/0