High Memory utilization on the CA-SDM Primary server

[SUSHB]

Dears,

 

We are facing memory utilization issue on the CA SDM primary server.(12.9). We are having 32 GB memory allocated for the Server(VM). it seems that server is using almost more than 90% of allocated memory. I am not sure if its Application issue or server issue.

 

Please let me know if any one having idea/suggestion about the issue. We are having only  CA SDM instance on this server, no other products on the server.

 

Additional Info:- We are having primary secondary configuration. secondary server is quite stable compare to primary server.(no high memory utilization)

 

 

 

Regards,

Sushant

[Jon_Israel]

Hi Sushant,

When looking at Task Manager on the server, can you tell us what is using most of the memory?  Which process or processes?  You can sort the list by memory usage and share a screenshot here if that helps.

Thanks,

Jon I.

[SUSHB]

Hi Jon,

 

Thanks for the reply. Here is the snap of memory utilization.

 

 

 

Regards,

Sushant Bansode

Givaudan GSD Support

 

On Tue, Aug 23, 2016 at 3:33 PM, Jon_Israel <

[Jon_Israel]

Hi Sushant,

So it is your webengines that are utilizing a lot of memory.  This could be due to a memory leak or that they are just caching a ton of data.  What I would recommend on this one is to opena  support case so that one of us can help you to gather some logs and string/object dumps to investigate why the webengines are utilizing such high memory.

Thanks,

Jon I.

[SUSHB]

Hi Jon,

 

CA support case already opened for this. Please go through if you have

access to view it.

00333404: CA Service Desk server not reachable.

 

last update from CA is to apply patch on the system. We are waiting for

down time so that i can apply this patch.

 

Its quite unclear for me if this is application issue or server issue. as I

can see on weekends when there are less number of users still memory

utilization is high. it gets down during weekly restarts. it builds rapidly

again during weekdays.

 

Regards,

Sushant Bansode

 

 

On Tue, Aug 23, 2016 at 3:58 PM, Jon_Israel <

[Vinod Subburaj]

Hi Sushant,

Before we start the troubleshooting, please ensure you applied the latest cumulative patch.

run pdm_webstat to find the list of logged in users.

pdm_vdbinfo should give the queued queries.

from the nx.env check if monitor join is uninstalled.

~Vinod

[SUSHB]

Hi Vinod,

 

Please find attached files for required info.

In nx.env file entry is :-

@NX_MONITOR_JOINS=No

 

Regards,

Sushant Bansode

Givaudan GSD Support

 

On Tue, Aug 23, 2016 at 4:45 PM, subvi02 <

Attachment(s)

pdm_vdbinfo_out.txt.zip   7 KB 1 version

pdm_webstat_out.txt.zip   425 B 1 version

[System]

Hi SUSHB - Did gbruneau's response help answer your question? If so please mark as Correct Answer. Thanks!

[Grant Bruneau]

We had a similar issue a year or so ago.  Using process explorer is was able to determine the VMware tools nic driver wasn't releasing memory.  I had our VM guy update VM tools on all our servers and that resolved the problem.

 

You will likely not be able to use Task Manager to troubleshoot as it only shows the private working set memory.  Process Explorer (link below) will allow you to see the Private Bytes used by each process (the amount of memory allocated to a process but not necessarily in use).

 

Process Explorer

[JEROME MAYER]

Hello Shushant,

Did you come to a solution to your problem?

/J

[SUSHB]

Hello,

 

issue is not resolved yet, as we are still waiting for downtime of

production server to apply patches suggested by CA. May be till 20 th

September things will be get more clear.

 

Regards,

Sushant Bansode

Givaudan GSD Support

 

On Thu, Sep 1, 2016 at 9:31 AM, jmayer <

[JEROME MAYER]

Thanks for the update.

Good luck then!

[SUSHB]

Hi Jerome,

 

Sorry for late reply. Just to update you, even after applying patches it didn't help us. Now sending interv logs to CA.

However i have noticed that even after stopping CA servicedesk services on this server, it is consuming around 74% memory, not sure if this is normal memory consumption. 

 

Regards,

Sushant

[JEROME MAYER]

HI Sush,

in fact if your OS already consuming 74% without SDM running, problem look to come from somewhere else and the remaining memory will mostly not be sufficient to run SDM.

One thing to check is what process /apps is consuming this memory and determine if the consumption is legitimate or not.

May you find orphan process like javaw.exe and you don't run any web server on that machine outside of SDM you may want to kill it to release that memory. I see time to time that those process are not killed correctly by SDM resulting on those orphan.

 

My 2 cents.

/J

[SUSHB]

Hi Jerome,

 

We are having primary & secondary configuration with 4 webengines on each

server.

as you mentioned that could be the case but not sure how to deal with it.

We are having weekly restart of the server. within a day or two again

memory consumption reach up to almost 96%. Do you have any idea how to fix

it permanently.

 

 

Regards,

Sushant Bansode

Givaudan GSD Support

 

On Wed, Dec 21, 2016 at 9:19 AM, jmayer <