disk extend operations fails (msg.disklib.NOSPACE) even though there is plenty of space on the datastore

[nthatte]

I am running ESXi 6.0.0 u3

I was extending disks on two datastores and it kept failing with the following error: The disk extend operation failed: msg.disklib.NOSPACE

I was logged into vSphere web and thick client trying to extend the disks and it was failing on both. I confirmed multiple times that the datastore had enough space and it showed it had plenty of space available. I ended up logging into the hosts directly that were hosting the VMs and I saw that hosts were not seeing the full size of the datastore. I had recently extending the datastore on my storage controller and increased the space on the LUN in vSphere. When I logged into the host directly, I saw that they were still seeing the old size. I refreshed and they hosts started seeing the new size and I was able to extend the disks on the VMs in vSphere web or thick client after that.

My question is, will this permanently update the hosts knowledge of the datastore size? Could they go back to thinking the datastore are smaller and shut the VMs down?

[a_p_ · Accepted solution]

I've experience similar issues with older vSphere 6.x versions too. Some of the hosts apparently recognized the new datastore size, but some of them didn't, which resulted in an inconsistent datastore size in vCenter Server (and on some ESXi hosts), depending on the host which last reported the size to vCenter. Neither rescans from the vCenter UI, nor the ESXi UI helped. The only reliable way to work around this issue, was to run vmkfstools -V (please note the upper-case -V) on each host. Once done everything showed up as expected.

André

[nthatte]

I ended up logging into each host via the vsphere client and refreshing the storage. That did the ttrick. I opened a case with VMware and they said I should be fine after doing what I did. As long as all the hosts see the new size, there should be no issues going forward.