October 6, 20178 yr Here are some notes to follow when trying to determine if issues with Virtual Servers on a vCMP have anything to do with not enough memory assigned to the vCMP guest. First of all don't rely on the GUI CPU & Memory stats as they are not very accurate. tmsh show sys cpu | head Sys::System CPU Information ------------------------------------------------------------------- System CPU Usage(%) Current Average Max(since 10/05/17 13:54:51) ------------------------------------------------------------------- Utilization 51 60 72 ----------------------------------------------------------------- Sys::Host CPUs ----------------------------------------------------------------- also run the following command to look at pagemem tmctl -d blade tmm/pagemem name used avail ------- ------ ------- pagemem 916182 1370624 Now calculate memory utilization percentage by doing some easy math %mem utilization = 916182/1370624 * 100 = 66.84% anything above 60%, we should start tweaking cores to be safer. Anything above 80% packet drops are seen and we’re in dangerous situation like 10/3. Here is how calculation for CPU & memory assigned for a particular VCMP guest based of the VCMP host configuration. Total CPU = number cores assigned for VCMP guest Total Memory = no.of cores * no.of slots assigned for VCMP guest * 2.81GB
Create an account or sign in to comment