I have recently started working for a new organisation.
The have a Hyper-V server (HP ProLiant DL580 G7, 4x10 core CPUs, 128 GB RAM) running few guest servers that do very little, but one of them is the primary SCCM 2012 server (32GB allocated), and performance is very poor.
We recently added WSUS to this server, and the performance is noticeably worse.
Performance monitor has just shown me pages/s peaking at 16,500, with an average of 250!
This is since we doubled the RAM last night and set SQL to use between 16-24GB ram.
Can anyone confirm if the pages/s is definitely indicative of insufficient RAM, as many articles i have read state?
There is sufficient capacity on the Hyper-V to double the RAM on the SCCM guest, but that means downtime and changes submitting.
Available memory is almost static, holding at above 12 gb free, no correlation with the pages/s spikes,
CPU is 40% average, min 16% min, 76% max - This is a little high i would say, and as the host has 40 cores, and this guest only 6, I am looking at upping them, but again this requires planned downtime.
The guest is likely to be quite loaded as it is a distribution point for SCCM. I would appreciate advice on which network counters should be monitored.
What else can i check/change?