Recently in my lab I ran out of diskpace. I was not able to clone VMs, nor power on VMs. In the past when this happened I had to Storage vMotion VMs off the storage, delete the VMFS partition, then delete the volume and create it again, format it as VMFS, and Storage vMotion the VMs back.
But I did not want to do that this time. I didn’t think I needed to - and I didn’t. No VMs were harmed or impacted during this very cool process. No outage!
This is what I did. I did this on NetApp HCI but it will likely work on other modern storage. Or at least maybe.
First, here are some of the errors I got:
The one above is an example of the error I got - a failure to power on a VM due to lack of disk space.
The one above occurred when I tried to clone a VM.
- So first thing to do is expand, or grow the source volume.
- So we log into the SolidFire UI, and change to Management / Volumes. Use the gear icon on the volume in question to access Edit.
- When you edit you have a number of options.
- We are going to change the volume size from 55 to 100. Just to make it obvious for us.
- Now we need to get VMware to see this change. We can see in the image below we have only 15.12 GB available.
- Now we right click on the datastore.
- Select the Increase Datastore Capacity option.
- It is hard to see, but it is the last option in the Select Device list. In the last column you can see a Y which means yes for expanding.
- Select the last option and press Next.
- I do not want to limit things so just hit Next.
- Now we can see the free space has not changed.
- We now right click on the Datacenter.
- Select Storage, followed by Rescan storage
- Next you will see a Rescan Storage dialog. Leave everything checked and hit OK.
- This will take a few minutes to finish, but when it does we will see new free space.
- There were several VMs on this datastore that did not crash or stop working while we did this. So pretty darn cool - right?
However, I did just break replication. The source and destination volumes are too different now. All that is needed is to access the destination volume and change its size to the new size of the source and a few minutes later it will be replicating again.
The error you will see for stopped replication in this case is PausedVolumeSizeMismatch. When you make the destination volume the proper size the error you will see is PausedDisconnected but it will only last 10 minutes or so.
Updates
- 3/17/20 - gave article a new title that makes more sense.
I hope this all makes sense, but if you have questions or comments let me know.
Michael
=== END ===
