We have a cluster of 3 SR850s spec'd up with 4x Intel 6242 CPUs and 1.5TB RAM on each configured as a cluster with low DRS and HA enabled.
We have a team who do regression runs between VMS so a SQL Server, App Server, Web Server and these run for hours on end sometimes causing CPU alarms of 80% upwards from them more so the APP and SQL Server.
This in turn causes DRS to get upset due to the mass resource hog and it then does its thing moving bits around to find a balance, however as this happens the regression testing slows and then in turn the next day stats get affected around overall run timings.
My question is, 1. do we have the right setup to cater for this sort of usage, 2. Would they be better on static hosts, 3. Would it be better to try and utilise the vApp service instead which treats the server bundle as a container ?
Your post needs moving, so I have reported it.