ESXi 4.1.0 servers loose Datastore after Starwind reboot

[drfurly]

I recently installed the free version of Starwind SAN (Version 5.8.0 Build 20120417) on a windows 2008 R2 machine. After setting up the Storage Server, I configured a Deduplication volume and a target for my esxi server, and was able to connect to the volume from ESXi and format it with VMFS. Before provisioning any Virtual machines, I wanted to test how the ESXi host would respond to a reboot of the Starwind Server or Service restart. As soon I restart the service or server, the Datastore greys out and never comes back online; and if I hit refresh the datastore disappears all together. I have to go through the process of Rescanning to see the volume again, and have to re-add the storage with the existing signatures. This worked the first three times before throwing a "hoststoragesystem.resolvemultipleunresolvedvmfsvolumes" message. At that point, I have to Assign a new signature to bring the datastore back online.

I've included the event log messages from the ESXi server. It looks like the disk ID is changing
First experience after a reboot.

Lost connectivity to storage device
eui.e5d089efa021afc1. Path vmhba34:C0:T0:L0 is down. Affected datastores: "SAN01-vBase-Core01".
Error 6/7/2012 3:21:19 AM

Lost connectivity to storage device
eui.5fa6e69c45ec25ec. Path vmhba34:C0:T0:L0 is down. Affected datastores:
"SAN01-vBase-Core01".
Error 6/8/2012 4:37:40 AM

Started receiving "hoststoragesystem.resolvemultipleunresolvedvmfsvolumes"

Lost connectivity to storage device
eui.08caf7ddf7e6c53f. Path vmhba34:C0:T0:L0 is down. Affected datastores: Unknown.
Error 6/8/2012 12:14:15 PM

Lost connectivity to storage device
eui.145ba9413a0fffb2. Path vmhba34:C0:T0:L0 is down. Affected datastores: Unknown.
Error 6/8/2012 12:16:44 PM


After re-signature the volume:
Just before:
Lost connectivity to storage device
eui.145ba9413a0fffb2. Path vmhba34:C0:T0:L0 is down. Affected datastores: Unknown.
Error 6/8/2012 12:28:02 PM

Just after:
Lost connectivity to storage device
eui.145ba9413a0fffb2. Path vmhba34:C1:T0:L0 is down. Affected datastores: "SAN01-CORE-FS1".
Error 6/8/2012 12:32:39 PM

Any help is appriciated.

Thanks,

Chris

[Anatoly (staff)]

First that I`d like to mention that it seams like this is pretty much VMware issue, not StarWind.

I think that StarWind logs would be more clear and useful for us, so would you be so kind to compress them and share with us?

Thank you

[drfurly]

Anatoly,

I'm not sure how you would come to the conclusion that this is a VMWare issue without any supporting data, but I am including the Starwind log as requested. Also, since experiencing this issue, I have been able to create a "image file" disk type that re-attaches without issue. Based on my testing the issue does not appear to be with ESXi, but seems to be related to Deduplication disk types. Please let me know if there is anything else I can do on my side to assist with getting to the root of this issue.

Thank you

[Anatoly (staff)]

It looks like the target was crated without the multisession - in the logs we see errors that related to the additional session. Would you be so kind to doublecheck this?

Also, DD device need some time for the "final prepare" after accomplishing the wizard. During this time it answers "device not ready, becoming ready" to the client requests.

[drfurly]

Hi,

I have recreated my test devices and enabled multisession on the targets, but I'm still experiencing the same behavior. Again, datastores based on image file device types come back online immediately everytime. I thought that maybe I was not giving the datastore enough time to recover after the simulated service outage, but even after 30 minutes the datastore would not automatically restore itself. However, if the ESXi server was rebooted, the datastore does automatically remount. Can you tell me why the image file backed datastore comes back online everytime, but the DD backed datastore does not? Is there any reason they would behave differently.

Thanks again for the assistance!

[Anatoly (staff)]

Would you please send your starwind logs to support@starwindsoftware.com?

[drfurly]

Enclosed, you will find my most recent log files. I'm trying to make a purchasing decision this week between Starwind and Datacore, so I'm hoping to get this resolved. We don't have this problem with the Datacore server we setup for testing against the same ESXi server config.

Also, I am unable to configure a Snapshot and CDP volume with an autocreation inteval (in minutes) with any value below 30 minutes. I have a customer requirement to provide snapshots in 15 minute intervals. Is this possible with your software?

Thanks

[Anatoly (staff)]

Our R&D reviwed the logs and currently we are working on the solution.

Stay tuned

UPDATE
Please check your mailbox - I have provided you with the link that should fix your issue. you just need to instal it over the old version.

[drfurly]

Anatoly,

Thank you very much for the new build. I will update my servers today and let you know how things turn out.

[Anatoly (staff)]

Great! We will wait for the update then!