Callmanager 7.1.5 - Low Available Virtual memory

[Roland Theisen]

Hello,

This morning some users complained that their Cisco CUCIMoc plugin is no longer working. When I wanted to login on the web page of my Callmanager, my Internet Explorer displayed me the error message: "Internet Explorer cannot display the webpage". So, I have checked my Real Time Monitoring Tool and I saw 2 errors:

1) Available virtual memory below 30 Percent. tomcat (951 MB) uses most of the memory. The alert is generated on Tue May 10 02:00:48 CEST 2011

2)Processor load over 90 Percent. tomcat (42 percent) uses most of the CPU.     Processor_Info:      For processor instance 1: %CPU= 98; %User= 68; %System= 7; %Nice= 21; %Idle= 2; %IOWait= 2; %softirq= 0; %irq= 0.      For processor instance _Total: %CPU= 95; %User= 59; %System= 9; %Nice= 26; %Idle= 5; %IOWait= 2; %softirq= 0; %irq= 0.      For processor instance 0: %CPU= 92; %User= 50; %System= 11; %Nice= 31; %Idle= 8; %IOWait= 1; %softirq= 0; %irq= 0.    The alert is generated on Tue May 10 02:06:18 CEST 2011 on node 172.16.0.11.       Memory_Info: %Mem Used= 74; %VM Used= 65.     Partition_Info:   Swap: %Disk Used=56.  Active: %Disk Used=43.  Common: %Disk Used=39.

Also the RTMT displays the message: "HTTP request failed. Web server unreachable."

I am running the version 7.1.5.32900 on my Callmanager and the firmware of the IP Phones is 9.0.3. My 2 servers are MCS 7825-H3 appliances.

I had this problem already two times since i made the upgrade to from the version 7.1.3 to the version 7.1.5. I could unblock the situation by rebooting the publisher server.

Can anybody help me?

Thanks a lot in advanced.

Marc Hoffmann

[Saurabh Agnihotri]

As an immediate relief from the solution , you may restart the Cisco Tomcat service and see if that helps getting things back up. You can do so by the following command from the CLI.

utils service restart Cisco Tomcat

I suspect that this is being caused by the common partition being full. Please attach the output of the following commands:

show status

show hardware

show version active

show process using-most memory

There are also many bugs around the same area but it would be difficult to point to a bug without the diagnostics from the server.

[hemasarode]

Hi,

CUCM : 7.1.5.34900-7

we have this same issue with on our publisher

"Available virtual memory below 30 Percent. DHCPMonitor (1163 MB) uses most of the memory. The alert is generated on Fri Jul 08 12:02:17 CEST 2016 on node XXXXXXXX."

We are not using CUCM as a DHCP server. So is it really necessary to have this feature service ACTIVE on our publisher?

What is an impact of restarting the DHCP Monitor service ?

I do not see any impact as such in our environment as of now but want to know whether this situation would lead to any issues?

We do not have a active contract with CISCO. so i am seeking a good advice here.

[shane.orr]

Were you able to resolve you issue?  I am having identical problems with 7.1.5.32019-1.

[Roland Theisen]

Hello,  For the moment I did not resolve the problem. I have perhaps a work-around which consists to switch off the RTMT. Our Cisco partner found a bug-fix in the latest relase notes of the CUCM v7.1.5. In this bug fix they speak about low virtual memory in combination with the RTMT running. So, for the moment my RTMT is switched off and my Callmanager is running fine.  But In the mean time I have ticket open at the Cisco TAC.   Best regards, Marc

[mudmathu]

Hi Marc,

Did you try restarting the Cisco Tomcat servie and did it help ?

You might be hitting a bug that is caused due to high no of http requests to the Tomcat by RTMT.

You can setup the AST sessions in the RTMT to monitor the memory  utilization and can set alert if the memory utilization crosses the  critical value.

Goto RTMT -> Performoance Monitoring -> Tomcat web-application -> Active sessions -> ast.

you will see the number of active sessions. Make sure that these AST sessions should not be more than 200-250.

You will gradually see that this value goes high due to memory leak. You can set the RTMT alert for this value as well.

Also please refer the TACEngineer this BUG :: CSCtj50884. Might be useful in resolving this issue.

Please do share your feedback when the issue is resolved.

Thanks

Mudit

PS: Rate if Useful

[Roland Theisen]

Hi Mudit, Yes, I had to restart the Tomcat service. Our Cisco partner already forwarded me a documnnt about the BUG::CSCtj50884. According to this document, the bug will be fixed to a futur version of the CUCM and due to a major VoIP project in my company, I intend to upgrade my Callmanagers to the latest version. But never the less, I will keep you informed about the status of this bug/problem. Thanks a lot. Best regards, Marc

[iptuser55]

i`ve seen this thoday as well as we are about to upgrade to 7.1.5 2b. It looks like it is fixed???? in SU4 - anyone done this?

[iptuser55]

Hi

I`m trying to down load the latest SU due to the above "UCSInstall_UCOS_UNRST_7.1.5.33900-10.sgn.iso" but when I try to download it it closes almost straight away - has anyone downloaded it recently

[Roland Theisen]

Hello,  Yes, it seems that this bug is fixed in 7.1.5 SU4. For the moment, I fine with the work-around which consists of switching off the RTMT.  Best regards,  Marc

[cooperben]

Not sure if this is relevant, but also please make sure to check your log tracing settings.  We had a similar situation--albeit with v6.1.5--and the high VituralMem/CPU usage was due to most of the debug traces being enabled for some reason.  Trimming those back really made a difference.  HTH

[Roland Theisen]

Hello,  I want just to inform you, that the Cisco TAC asked me to upgrade my Callmanager to the version 7.1.5 SU4. So I will organize now this upgrade in the coming weeks. I will post my results afterwards.  Best regards,  Marc Hoffmann