Capacity Remaining Calculations

[chan82]

Hi guys,

We are deploying VCOPS and configuring it.

Our environment consist of two ESXi clusters with two datatstores each. On top of that, there is a datastore (named data1) presented to both clusters for ISO files storage purpose.

It seems that datastore (data1) is duplicated in VCOPS as it's capacity is reflected on both clusters and it is affecting our capacity remaining calculations.

Is there a way to exclude that datastore from VCOPS capacity remaining calculation?

[mark_j]

You could apply a specific config policy to that datastore, which has a capacity buffer of 100%.. this will cause vC Ops to not consider it usable, plus the ISOs on it won't be considered VM disks and won't be considered used either.

[chan82]

Hi Mark,

Thanks for replying.

I have configured a specific policy for the datastore although VCOPS only allow me to set the buffer to max 99%. I associated the policy to the datastore. On datastore leve, I can see the effective policy is that specific policy.

On cluster level, I am still seeing the usable capacity as same as before creation of the policy. Am I missing something or it takes time for VCOPS to re-calculate?

[mark_j]

Is your config policy set to use "usable capacity" for the cap planning model? Check tab 3a in the config policy. Sometimes the disk space planning metrics can take a little while to recalc. The planning badges also typically re-calc every 12 hours or so.

[chan82]

Yes, we are using usable capacity for all of our policies. Is it why we are allowed to set buffer up to only 99%?

I will check again on the capacity planning portion.

Another question:

If buffer is set on the datastore, will it be reflected on the cluster level in terms of usable capacity or the cluster's usable capacity is calculated based the buffer set in the cluster associated policy?

[mark_j]

The buffers will be taken in to for usable capacity, so the policy that sets 99% buffer should leave only 1% usable, theoretically.

[chan82]

It seems that the on cluster level, the usable capacity still includes the datastore with the excluding policy. The cluster probably take the setting from the policy associated with it. I did not notice any changes in the capacity remaining score and no of available VMs remaining after one week.