Forum Discussion
More info:
We recently converted from some ageing physical appliances and since we went live we have seen an increase in response times from a web service that has a sub-second SLA. Mostly ping times to the web service from outside the LTM are around a millisecond, put periodically they'll climb to as much as 200-300 milliseconds. My first instinct was that we might have a CPU scheduling issue at the VMWare layer, so I set CPU reservations but that didn't impact the problem. During further investigation I ran across the discrepancy in the number of sockets. I plan on changing the config to 2 dual core virtual sockets during our next maintenance window, but I'm still curious if anyone has any idea about what performance would be like in current conditions.