Created attachment 889006 [details] Cluster Optimization in RHEV 3.4 Description of problem: Enhancements in 3.4 to Enable/Disable KSM have introduced confusion regarding terminology being used in the Cluster -> Optimization screen. Historically, we have had settings for "Memory Optimization" which provide options of None, For Server Load, and For Desktop Load", which apply various "memory page sharing" values. I believe "memory page sharing" is an improper term, as to me, and several customers, relate that to VMware's feature of "Transparent Page Sharing", which is equivilent to our KSM. Now that we have added the option to Enable/Disable KSM, this becomes confusing, as it seems like we can modify the same concept in two different locations. How reproducible: 100% in RHEV 3.4 Steps to Reproduce: 1. Create or Edit a Cluster 2. Click on the "Optimization" tab Actual results: As defined above Expected results: We should look to modify the terminology of the Memory Optimization setting to more accurately reflect what it is controlling, which, as far as I can tell, is Consumed Memory Over-Commit, or how much we can over-commit a cluster or host in the cluster based on the idea that it is possible to Allocate more memory to VMs that exist in the source host. Memory Page Sharing should not be used, as it causes confusion agains what KSM actually does... which is physical page sharing. Additional info: See BZ 1072030 for how this is already causing confusion in RHEV 3.3 without the KSM Control option
this bug status was moved to MODIFIED before engine vt5 was built, hence moving to on_qa, if this was mistake and the fix isn't in, please contact rhev-integ
Verified on - 3.5.0-0.15.beta.el6ev
rhev 3.5.0 was released. closing.