Datastores offline after SW reboot

craggy · Wed Jul 08, 2015 10:15 pm

We have had the same issue on 3 differet ESXi clusters connecting to 4 different Starwind servers whereby rebooting the SW server results in the datastores going offline and not connecting after the SW server has come back up. Rescans of the LUNs etc. don't help. The datastores are permanently offline until the ESXi hosts are rebooted.

However, we have the same 3 ESXi clusters connected to Nexenta storage or Solaris storage with Napp-it and as soon as they reboot the datastores reconnect immediately.

It's really messy that a reboot of Starwind whether expected or unplanned requires a reboot of all hosts to get the datastores back online.

What is the cause of this and is there any solution for it?

craggy · Fri Jul 10, 2015 11:54 am

Anyone??

darklight · Wed Jul 15, 2015 4:18 pm

Here Hste posted a nice idea for ESXi.

https://forums.starwindsoftware.com/vie ... 244#p24284

Take a look.

hste · Thu Jul 16, 2015 8:54 am

That was another scenario with esx reboot. I see this is after a starwind reboot.

I guess there is some locking issue. You could look at this kb http://kb.vmware.com/selfservice/micros ... Id=1004033 it might be the problem.
I know there has been some problems with vaai and san, you could try to turn off vaai to see if it occurs with vaai turned off. What version of esx do you use?

hste

craggy · Sun Jul 19, 2015 8:07 pm

I use ESXi 5.5 everywhere.

I just don't get this issue with any other SAN software so don't think it's a VMware issue.

Also, I doubt that script will help no amount of manul rescanning or trying to remount volumes from the cli will work.

darklight · Mon Jul 20, 2015 9:02 am

Do you reboot all the SW servers at the same time? Maybe they are just syncing after reboot and you should wait for some time to get the SW devices active again?

craggy · Wed Jul 22, 2015 4:36 pm

The servers are not in a HA set up, they are standalone storage servers as they have about 60TB storage each and cost of running HA is prohibitive.

darklight · Fri Jul 24, 2015 8:37 am

Since rebooting of ESXi hosts temporarily resolves the problem it seems to be a kind of VMWare mess I believe. Do you have all the critical updates installed?

Tue Jul 28, 2015 3:08 pm

Hi craggy,

Please provide me with the following:

type of StarWind image
type of L1 and L2 cache if enabled
is it HA or standalone image
StarWind build number