Help! I lost a whole partition/target

hennish · Thu Sep 03, 2009 5:39 pm

I am running a PoC for my customer's test environment using 2 VMware ESX servers and a Windows 2003 Server running Starwind Server.

I first set up a 50 GB "Image File" (name: san1) target on an existing partition, and tested it out. Worked fine. Put a few VMs on it. No problems.

Then I popped in a few more disks, created a RAID volume using HP Array Config Manager, and created a Disk Bridge Device (name: san2). At first, I had some problems connecting to it from my ESX servers, but after a reboot of the Windows server running Starwind, it showed up in the ESX servers, and everyting worked fine.

Now, suddenly four of my VMs are down, and I can't reach the san2 from the ESX servers, dispite trying a reboot of the Starwind server again.

The ESX servers can see the LUN, but can't access it. If i do a "fdisk -lu /dev/sdb", the shell just hangs.

The Starwind GUI says "san2" is "Not Active", but I can't find anywhere to activate it. The Starwind log says "Reject: target 'san2' device 'DiskBridge0' is not active" and "IN_LOGIN: Event - LOGIN_REJECT".

What to do? These servers are not mission critical, but it would be kinda bad to lose them completely.

Robert (staff) · Fri Sep 04, 2009 1:01 pm

hennish,

I think that Dmitry from our support department is already assisting you with this case.

I will provide an update here for everyone else once we have the details.

Thanks

hennish · Tue Oct 06, 2009 11:41 am

Here is a late reply: The problems were caused by Windows somehow changing the internal LUN ID of the disk device in question. I got assistance from the support crew to simply recreate the DiskBridge device, and then everything jumped into place and started fine again.

Robert (staff) · Wed Oct 07, 2009 9:43 am

Glad that everything is fine now!

Thanks

kdsnice · Tue Oct 20, 2009 12:58 pm

Robert (staff) wrote:Glad that everything is fine now!

Yes, works after recreating iSCSI target with same name. But this situation repeated every time after reboot server with StarWind.

Robert (staff) · Thu Oct 22, 2009 8:30 am

That may be related to ESX itself. Try to change LVMresignature parameter to "1" in the advanced options of Vmware. This could help.

Thanks.