DRS on a cluster with two ESX Hosts is set to Fully Automated mode. One server in the cluster is running at 98% CPU utilization, while the other host is running at 52% utilization.
Upon analysis, it appears that none of the virtual machines (VMs) are being migrated to the other server, and no recommendations are being generated. Which of the following are the most likely causes of the problem (Choose Two)?
A.
The ESX Server has CPU features that are incompatible with other hosts in the cluster
B.
The VMotion licenses for the cluster have expired
C.
The current DRS Migration Threshold setting is too conservative
D.
Insufficient resources exist to perform the VMotion migration
Explanation:
B- is not correct, there is no VMotion license
C- is incorrect becuase they analyzed and none of the machines are moving
A – a posibilty for sure
D- RAM maybe unavailable, not just CPU
I believe that C&D are the correct answers. In DRS the there is a slider bar that can be moved from Conservative to Aggressive. If the slider bar is all the way to the conservative side that would explain no recommendations being generated there by not vMotioning any VM’s. I have had personal experience with this option. Also, wouldn’t a DRS cluster require all the CPUs on ESX hosts in the cluster to be of the same family there by making them compatible? This would rule out A.
Won’t you still get recommendations to migrate if the threshold was the issue?
“Also, wouldn’t a DRS cluster require all the CPUs on ESX hosts in the cluster to be of the same family there by making them compatible? This would rule out A.”
I think only if EVC is enabled right?
If you put threshold at 1, that is “conservative”, the DRS does not do any recommendation except for contraints like host in maintenance mode or anti-affinity rules. So I think “C” is correct! Moreover they have seen “there are no recommendation” then it is the DRS alghorithm that did decision of not making recommendation.
I go with C and D
C – no migrations and recommendations are happening because the setting isn’t aggressive enough
D – RAM on the other machine could be a factor, we’re only given the CPU figurs.
A and B are wrong.
I too would go for C and D.
C is definetly correct. Priority 1 say “vCenter will only apply recommendations that must be taken to satisfy cluster contraints like affinity rules and host maintenance”
The second could be A or D, Iam unsure. Cause EVC Mode must not be enabled in Cluster. Therefore we could have 2 AMD or Intel hosts but maybe with different CPU Features, therefore this could be possible. D could also be. In ESX 3.5 B was also possible.. Strange question.. I think there’s somethign missing/wrong in the question…
I’d think that only C is correct. If A or D were correct recommendations would be generated and then (I’d expect) a warning or error would be generated. Does anybody have a lab to test?