3Par 7200 - Storage Issues - Zero Detect

[SimonGreen001]

Hi All,

Could do with some insight into an issue I am facing. 

Long and short is that I'm seeing a significant difference between VMWare datastore consumption and underlying tpvv on a couple of volumes. As an exmaple one datastore has a single vmdk of 600GB and the tpvv is showing 1716GB total used. 

I'm facing storage capacity issues and after some digging I can see that the tpvv's have been set with Zero Detection disabled. I'm starting to think that this is the underlying issue here?

Is this a setting that can be changed 'on-the-fly' for a tpvv and any gotchas?

Appreciate any insight as it's been a long and painful couple of days

Simon

[Cali]

Hi Simon,

I assume your Datastore is VMFS-5 formatted.

In this case, ESX doesn't notify the Storage about unused Blocks and the Volume on the Storage side never shrinks.

Solution 1 (For long Term):

If you are on ESX 6.5 or higher, reformat the Datastore with VMFS-6 (Move the Data, you will lose it).
See: Versions of VMFS Datastores (vmware.com)
-> Automatic space reclamation

Solution 2 (Short Term):

Use the following command for every Datastore on one of the ESX Hosts:
esxcli storage vmfs unmap -l Name_of_your_Datastore -n 6400

You will see nothing on the CLI for 1/2-1h before the CLI prompt come back and there is no success Message, but this is OK.
Watch the Volume Size on 3PAR, it will shrink continuously during this Time.

Cali

[SimonGreen001]

Hi Cali,

Really appreciate your response.

The VMSF datastores are formatted VMFS-6..

I've checked the space reclaimation settings on the datastores this morning and it was set to low. I then ramped it up to 2000mb/s all to no avail. I'm simply not seeing any reduction on the tpvv's at all. 

If I look at VMware datastore 1 I can see a 1tb datastore with a single 600mb VMDK. If I look at the underlying tpvv I can see 183% user used, virtual size 1024gb and reserved size of 1.873tb. We have no more raw capacity and hence CPG group SD space allocation failure again.

Any insights appreciated.

Simon

Any ideas appreciated..

Si

[Cali · Accepted solution]

Hmm, mow it's tricky.

ZeroDetect many safe some Space but not so mutch.

Try to run the escli Command, if there is nothing to do, it will run only for minutes.

Do you run Compaqt CPG?

The 7200 do very slow Space Reclamation, check with "srvvspace -hourly -btsecs -12h", look for "Rsvd", is it going down last 12h?

If yes, Storage is still recovering free Space, as far as I know, there is no User Way to Speed up (Support can do).

The only way I know to free up Space immediately on the Storage is to delete the Volume.
If you have 600GB of Space free, move the VM to a different Datastore, unmount and delete the Volume.