VMs Hard drive in Read-only state after Storage Fail-over

[marcinmf]

Hello, 

We had an issue today where our storage triggered fail lover between controllers.

During that process our Linux based VMs went in to read-only drive state to avoid data loss or corruption.

When storage recovered we had to reboot Linux VMs to get out from read-only state, 

CentOS systems rebooted with out issue - repaired itself and booted to normal read write drive state. 

Ubuntu VMs had to be re scanned after boot to come back online and few did not boot up at all with swap partition missing / not able to mount.

Windows VMs were not affected.

Here is our setup:

NetApp Storage FAS8040 - dual controller configured for HA running Ontap 9.2.

ESXi - 6.0.0, 7967664 with multi-path connection to storage via ISCSI

Ubuntu  16.04.3 LTS  xenial - Disk Time out set to 180 - no multi path connectivity.

Is there specific Ubuntu setting for vmware that we are missing? Or maybe ESXi need specific configuration for NetApp?

From what i was able to find online ESXi before 6 were having this issue and Linux with older kernel but it should be patched with what we have right now.

I appreciate any suggestions.

Martin

[andris]

It's always good to check for compatibility and any caveats in the IMT.

The NetApp® Interoperability Matrix Tool (IMT) defines the qualified components and versions you can use to build FC/FCoE, iSCSI, NFS and CIFS configurations.

Note: ONTAP 9.2 is no longer actively supported by Engineering  -- it's in a limited support state until 2020.  If your environment allows (check IMT), consider upgrading to ONTAP 9.5P6 -- it's the cat's meow, currently. If you have IOM3/DS4243 storage shelves, you'd have to stick with 9.3Px -- 9.3P15 is the latest/greatest.