My main datastore became read-only...

[jspiel]

I have an issue on my ESXi 6.0 (up to date, with last update ESXi-6.0.0-20160504001-standard ).

This morning, the 2 VMs of my personal ESXi server were powered down.

I connected to the esxi host and tried to power them on without success.

When doing some tests, i saw that the datastore was read-only, i can't even create a folder or do a simple touch. But, i can delete files...

[root@esxisrv:/vmfs/volumes/54575f35-e5e43590-241f-3c075465b1dd] ls mosx/

archives      tmp            xsibackup.NEW

testbck        xsibackup      xsibackup.old

[root@esxisrv:/vmfs/volumes/54575f35-e5e43590-241f-3c075465b1dd] rm -rf mosx/tmp

/

[root@esxisrv:/vmfs/volumes/54575f35-e5e43590-241f-3c075465b1dd] touch toto

touch: toto: Input/output error

[root@esxisrv:/vmfs/volumes/54575f35-e5e43590-241f-3c075465b1dd] ls mosx

archives      testbck        xsibackup      xsibackup.NEW  xsibackup.old

[root@esxisrv:/vmfs/volumes/54575f35-e5e43590-241f-3c075465b1dd]

I tried to reboot, and it didn't change anything. I'm now copying the VMs from a backup to my second datastore to launch them, but that won't help me to understand and fix the issue.

I checked SMART status and it says it's ok. I have no idea where to look now.

[continuum]

Can you copy/extract your VMs or do you see I/O errors when trying ?

[jspiel]

Can you copy/extract your VMs or do you see I/O errors when trying ?

Yes i was able to copy one of the VMs from this datastore to the second one and it worked like a charm on the new datastore. (and there was no error)

For the second VM, i copied from my backup.

It's really only a problem when trying to write on it.

[Viewaskew]

You say the VMs were powered down? I assume not but best to ask, were there any changes in config or software whatsoever ?

[jspiel]

Viewaskew‌ said :

"You say the VMs were powered down? I assume not but best to ask, were there any changes in config or software whatsoever ?"

I'm using xsibackup for my backups, and it starts by doing a snapshot. I suppose the VM went down when it tried to do the snapshot and realized the datastore was not writable. But that's only a supposition. By looking at the last modified time of the VM files, it looks like the 2 VMs were not shut down at the same time. That also goes in the way of the powering down when trying to backup it.

And no, no config change or software update...

[Marmotte94]

Hi,

Can you check if you are the same problem on an other ESXi Host.

Can you check on you storage if you can set your initiator to RO for your ESXi Host.

Thank you,

Regards,

[jspiel]

Can you check if you are the same problem on an other ESXi Host.

Can you check on you storage if you can set your initiator to RO for your ESXi Host.

The datastore is a local SSD disk. (actually the same as where ESXi is installed)

And the SMART Health status is OK.

[jspiel]

Any idea where i could look for this issue ? Performances of my second datastore are really low compared to the internal SSD, and i'd love to move back my 2 VMs to the SSD datastore.