Problem with vMware HA Cluster resource reservation values

[Huolsam]

Hi,

I have a problem with vMware HA Cluster resource reservation. The cluster total capacity is 13.97 Ghz but only 7.82 Ghz is available.

I didn't understand why, if someone could explain it to me.

Also "Total slots in cluster" calculation is based on the 7.82 Ghz ( 7.82 * 1000 / 32 = 244 )

im using vMware vCenter version 6.5.

thanks.

[nachogonzalez]

hey, hope that works:https://kb.vmware.com/s/article/76854

VMware Knowledge Base

[scott28tt]

Moderator: Thread moved to the vSphere area.

[Huolsam]

Hi nachogonzalez​,

The KB talk about "The Total Reservation Capacity value is larger than the Cluster Total Capacity", but in my case The Total Reservation Capacity is half of the Cluster Total Capacity.

[depping · Accepted solution]

You configured admission control it looks like. Considering you only have 2 hosts, and the default is to take 1 host failure in to account, it basically means you "lose 50%" capacity as that capacity is being reserved in case 1 of the hosts fails and the VMs on that host need to be restarted.

[Huolsam]

Thanks depping​ for the answer.

[MVL_Lopez]

Hi, this is completely true, however, I have the same problem here, but I have not enable HA in the Cluster, so there is no Admission Control. And now? Why the Total Reservation Capacity is half the Cluster Capacity?

[vbondzio]

If that is indeed the case, then most likely because the hosts are tiny and ESXi is reserving ~2 cores worth of CPU for vim (hostd, vpxa, etc.). Note that "reserved" doesn't mean used, just that ESXi "reserves the right" to use the CPU whenever it might need to.

To verify, run this on the hosts and post the output either on e.g. pastebin or in pre tags (like the command) back here.

sched-stats -t groups -z vmgid:name:pgid:pname:size:vsmps:amin:amax:units:resvMHz:availMHz | awk 'NR == 1 || $2 ~ /^(vm\.|pool)[0-9]+/ || /^[0-4] / {print $0}'

[MVL_Lopez]

Hi, this is the result of the command in one of the cluster hosts:

[vbondzio]

Jup, those host only have 4 cores so the overall reserveable capacity is low. The system group is reserving 10% of a core and the vim group 1461 MHz, that is normal.

[MVL_Lopez]

Thanks you, do you have a similar command to analyze the memory?

[vbondzio]

A similar command for memory would e.g. be:

memstats -r group-stats -g0 -l2 -s gid:name:parGid:nChild:min:max:conResv:availResv:memSize -u mb 2> /dev/null | sed -n '/^-\+/,/.*\n/p'

I attached an exploded pdf of a presentation that goes through some of this.

[MVL_Lopez]

Excellet PDF. Thank you.