IBM Verify

IBM Verify

Join this online user group to communicate across Security product users and IBM experts by sharing advice and best practices with peers and staying up to date regarding product enhancements.

 View Only
  • 1.  how manually spike memory of ISAM appliance?

    Posted Tue April 06, 2021 09:42 AM
    I am trying to test our monitoring device that polls snmp traps from iSAM appliance for monitoring memory metrics. In our test servers I can't see the memory even reaching 1% and the minimum threshold I can set in my monitoring tool is about 1% hence I am not able to achieve an alert to test. so checking if you have any ideas on how to trigger a memory spike on the ISAM appliance. Thank  you!

    -Raj.

    ------------------------------
    Rajkumar
    ------------------------------


  • 2.  RE: how manually spike memory of ISAM appliance?

    Posted Tue April 06, 2021 04:17 PM
    Raj,
     
    If the appliance is running a WebSEAL server the easiest way to consume more memory is to simply create many authenticated sessions (without performing a logout).  Alternatively, each WebSEAL instance consumes a certain amount of memory and so you could temporarily create a large number of WebSEAL instances.
     
    I hope that this helps,
     
    Scott.
     
     
     

    Scott A. Exton
    Senior Software Engineer
    Chief Programmer - IBM Security Verify Access

    IBM Master Inventor

     
     





  • 3.  RE: how manually spike memory of ISAM appliance?

    Posted Tue April 06, 2021 04:26 PM
    Thanks for the ideas Scott.. you are best!!





  • 4.  RE: how manually spike memory of ISAM appliance?

    Posted Tue April 06, 2021 04:48 PM
    You could also increase the number of threads ([server] worker-thread) from 300 (default) to a large number and it should increase the memory footprint.

    ------------------------------
    Sylvain Gilbert
    ------------------------------



  • 5.  RE: how manually spike memory of ISAM appliance?

    Posted Wed April 07, 2021 07:58 AM
    Thank you Sylvain for you inputs.., I'll try this as as well. I understand that additional worker threads consume memory, roughly about 512KB each. Is this the concept behind that? I have a quick question regarding this concept, if I simply increase the global worker threads to 1000 and I don't create any junctions on the instance..is it still going to allocate memory to that instance? I thought the memory utilization is also proportional to the number of junctions on that instance. Sorry, I deviated from the actual subject. 





  • 6.  RE: how manually spike memory of ISAM appliance?

    Posted Wed April 07, 2021 09:00 AM
    I changed in one system worker-threads from default 300 to 30000 (x100) on an exisiting Web Reverse Proxy and I could see an increase but maybe not that huge one you're looking for. Without putting any trafic on that Web Reverse Proxy, I could see the overall appliance memory usage increase from 2.77 GB to 3,84 GB (value picked up as well with SNMP Monitoring configured) after the WRP restart (no Appliance reboot). This particular Web Reverse Proxy had no junction defined at all. The LMI was really slow after the Web Reverse Proxy was restarted confirming the resource grap that was operated by the worker-threads change, before stabilizing.

    In the LMI, in the URI /core/analysis/system_memory page, Memory Used displaysed is almost 50% whereas before the worker-threads change it was barely consuming 1% (of total 4 GB of RAM assigned in the Appliance).

    You could continue to crank that worker-threads value further to see what happens.

    Hope this helps.

    ------------------------------
    Sylvain Gilbert
    ------------------------------



  • 7.  RE: how manually spike memory of ISAM appliance?

    Posted Wed April 21, 2021 07:43 AM
    Sylvain, Thanks for your inputs , I have tried what you said and I have seen exactly the same behaviour. I did notice the CPU hitting 90% when I restarted the instance later it calmed down. The memory was elevated as long as I used 30K threads and it reduced when I reduced the threads back to 300., This is a good experiment. I also tried adding multiple instances like Scott said and I did see a memory spike. In a way both of them are basically increasing worker threads and they have a direct correlation to Memory. Thank you both for your help. 
    -Raj.