Order of action

[l0ad3dr3v0lv3r]

Hi all,

I'm new to vSphere and have the following setup:

4 HPE Gen8 servers running ESXi 5.5. Theres is additionally a  SAN connected to these as their datastore.

Recently I was tasked with moving all the VM's on two of these servers and consolidating them on the remaining two servers which I have done successfully. I've also configured the VM's to use local storage as their datastore freeing up the SAN.

The idea here is that they want to pull the two servers and SAN that are now doing nothing out of the rack and use them elsewhere. However, they've also asked me to ensure that all the disks in the SAN have been wiped clean before anything else.

I see that I have a datastore cluster that was configure on this SAN and was wondering if there is any order that I need to follow before blanking the disks (HDD). I've read that I can use the CLI to blank the disks with the following:

# dd if=/dev/zero of=/[mounted-volume]/zeroes && rm -f /[mounted-volume]/zeroes

Is there anything I'm missing? I'm sure there's an order to the steps that I need to preform.

I do appologize in advance if I've missed any critical information or the above doesn't seem to sound right, again, I reiterate....I'm new to vSphere.

Thanks all

[daphnissov]

One second before going there. You say:

Recently I was tasked with moving all the VM's on two of these servers and consolidating them on the remaining two servers which I have done successfully. I've also configured the VM's to use local storage as their datastore freeing up the SAN.

Does this imply that the VMs on what was a 4-node cluster were on SAN-based storage and are being moved to local storage on 2 nodes? If that's correct...why?

[l0ad3dr3v0lv3r]

Hi daphnissov,

Thanks for taking the time to reply. Yes, the four servers were in a 4 host cluster and the VM's were sitting on the SAN. The reason is because it was decided that the VM's were not growing as anticipated and that the SAN could be used elsewhere. Additionally, the cluster of 4 hosts was overkill and unnceccessary for the enivorment (just a DEV enviro) that they were being utilised for.

So, the idea was to free up the SAN and free up to of the servers. Consolidate the VM's on the two remining servers and away we go....

[daphnissov]

Sure, I understand that, but by moving away from shared storage and onto host-local storage brings implications that could be dangerous. Do you have a replication or backup strategy in place that will protect these VMs that are now sitting on local storage in case a host failure? In case logical corruption? Etc?

---

That aside, once you've vMotioned all your VMs onto local storage, you need to unmount and then delete those VMFS volumes (assuming it's block). See one of my blog posts to know how to do that correctly. Then, check with your storage array vendor on the re-initialization process. Many vendors have a wiping tool they recommend (or that's built into the microcode itself). Using a disk wiping tool from within ESXi is not sufficient to do this as there could be data elsewhere on those array disks.

[l0ad3dr3v0lv3r]

Thanks again.

So, I guess the only thing remaining is the datastore cluster. Unfortunately, the SAN doesn't live alone. When this was setup, a datastore cluster was built with the SAN and several other storage sources to make up the cluster. I'm assuming I'll need to get rid of the cluster before I'm able to do anything with the physical disks in the SAN? I.e Unmount and delete the VMFS volumes?

[daphnissov]

A datastore cluster is a logical construct--it doesn't exist as a physical entity on ESXi hosts. You'll want to follow the same process on each extent (datastore) which comprised the datastore cluster. In cases where some of those extents aren't backed by an Enterprise-grade storage array, you can try a disk-wiping tool (making sure you know exactly what you're doing) but would need to be done from a client who has access to that extent.