Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

IRP VM Memory Utilization is high 80% continuously: - CWMS 1.5 MR1

Hi Team,

We have installed CWMS on premise server with HA.

It contains following set of Vm’s, Every 2nd VM is HA.

2 Admin

2 Media

2 IRP

Last week the IRP service/Vm was failover to HA IRP Vm but since 29th Dec 13 the HA IRP VM memrory utilization is showing high & it is 80% continuously.

However the meeting frequency is between10-20 only and we have 800 concurrent USERS environment of cwms.

Can help us to understand the cause?

Does memory utilization will get reduced?

Thanks in advance.

1 ACCEPTED SOLUTION

Accepted Solutions
Cisco Employee

IRP VM Memory Utilization is high 80% continuously: - CWMS 1.5 M

Hi Pratik,

Memory utilization in the 70+% range  is normal since the Linux system will attempt to cache as much of the active data into RAM as possible, so even when idle memory utilization will be high on the VM. The least used part of this cache will be released on demand as other files need to be loaded into RAM, but otherwise the system will leave the data in active RAM so it can be accessed quickly. This is normal behavior to be the same as any Linux system, reserving as much memory as possible. 

The dashboard only shows the memory usage on the VM with the highest amount of utilization. This is usually an IRP Node because there is not as much memory assigned to those VMs. As for the alarms, since this is expected for your system I suggest increasing the memory 'limit' for sending alerts to 90%. It is unlikely that the system will stay at those memory levels for multiple hours unless there is something wrong or stuck. Most of this memory is not actually actively being used or locked. The system will release this on demand when large chunks are needed or if the cached data is not needed for a long time.

Furthermore, since this is normal for almost all systems deployed, this is being redesigned in version 2.0 to report memory usage properly.

-Dejan

9 REPLIES
Cisco Employee

IRP VM Memory Utilization is high 80% continuously: - CWMS 1.5 M

Hi Pratik,

Memory utilization in the 70+% range  is normal since the Linux system will attempt to cache as much of the active data into RAM as possible, so even when idle memory utilization will be high on the VM. The least used part of this cache will be released on demand as other files need to be loaded into RAM, but otherwise the system will leave the data in active RAM so it can be accessed quickly. This is normal behavior to be the same as any Linux system, reserving as much memory as possible. 

The dashboard only shows the memory usage on the VM with the highest amount of utilization. This is usually an IRP Node because there is not as much memory assigned to those VMs. As for the alarms, since this is expected for your system I suggest increasing the memory 'limit' for sending alerts to 90%. It is unlikely that the system will stay at those memory levels for multiple hours unless there is something wrong or stuck. Most of this memory is not actually actively being used or locked. The system will release this on demand when large chunks are needed or if the cached data is not needed for a long time.

Furthermore, since this is normal for almost all systems deployed, this is being redesigned in version 2.0 to report memory usage properly.

-Dejan

New Member

IRP VM Memory Utilization is high 80% continuously: - CWMS 1.5 M

Thank you very much for explaining & sharing this information.

It help us to understand CWMS as well as backend.

Thanks Again!

New Member

Hi , Good Day. We are using

Hi ,

 

Good Day.

 

We are using the CWMS 2.0 MR3 .

Base on your statement . May i know if the version 2.0 able to provide the memory usage correctly ? Will the linux cache behavior still causing the memory usage reporting high (>70%) ?

"

Furthermore, since this is normal for almost all systems deployed, this is being redesigned in version 2.0 to report memory usage properly.

"

Thanks.

Cisco Employee

Hi, In CWMS 2.0 memory

Hi,

 

In CWMS 2.0 memory utilization is also presented based on the VM that has the highest utilization on the system. Because of this design limitation, in version 2.5 that is coming out soon, we are completely removing CPU and RAM meters from the dashboard and will document the ways to properly monitor resource utilization using vCenter tools.

Kind regards,

 

Dejan

 

 

New Member

Thanks for the explanation .

Thanks for the explanation .

New Member

Hello Dejean,We have

Hello Dejean,

We have installed CWMS 2.0 MR6 2000 ports, and we have intermittent high memory usage alarms on some servers (76%).

We would like to know which alarm  threshold should be set (75% or higher) to correctly monitor the system health ?.

In your previous post you are talking about a document to properly monitor resource utilization using Vcenter. Is this document is available today ?

Regards,

Marc

 

 

Cisco Employee

Hi Marc,Brief spikes in

Hi Marc,

Brief spikes in memory usage are fine, but if your memory usage stays continuously at such high levels and keeps increasing, you may be hitting a bug of some sort. In 2.0 MR7 there is a fix for several performance bugs causing higher CPU utilization, so if you want, without doing any analysis of your system, and update to 2.0 MR7. After that, in case you still experience high memory usage, I would advise opening a ticket with TAC to look into this.

 

-Dejan

New Member

In your previous post you are

In your previous post you are talking about a document to properly monitor resource utilization using Vcenter. Is this document is available today ?

Cisco Employee

Hi Marc, There isn't a Cisco

Hi Marc,

 

There isn't a Cisco document available for that because as you know all this is running on VMware platform. VMware has guides for monitoring performance of objects in vSphere client (vCenter). Each version has it's own, but here is the link to vSphere 5.1: http://pubs.vmware.com/vsphere-51/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-51-monitoring-performance-guide.pdf

I hope this will help you.

-Dejan

665
Views
20
Helpful
9
Replies