Fred Schmidt <[EMAIL PROTECTED]> writes: >However, I would like to persist with CMM. My original question still >stands... how can I verify that it is actually working? >
>From the z/VM perspective you should see resident pages and total pages for virtual machines being managed drop when the system comes under constraint. For Performance Toolkit, this would be seen in UPAGE report (FCX113). The process of Linux freeing backing memory is done via Diagnose x'10'. You can see the diagnose 10 counts in PRIVOP report (FCX104). Some other items of interest may be users in page wait, see USTAT (FCX114). Other performance products will have similar reports. >From the Linux perspective, you can check the size of the pool of pages put out of play via the commands: cat /proc/sys/vm/cmm_pages cat /proc/sys/vm/cmm_timed_pages Out of the box, VMRM is configured to use the first pool. Regards, Bill