SATA Disks bigger then 160GB

[Mitschele]

Dear Community,

i am working with Starwind Software for a few Month now and sofar everything has worked fine. I use Starwind on a iSCSI-SAN in a VMware ESX3.x environment without any Problems up to now. I have setup 2 * 160 GB Drives as Bridge Devices to the ESX Servers and they contain vmfs3 Filestructure to contain Virtual Machine Files. Last week i have replaced one of the 160GB Bridged Disks with a 500 GB Disk. All went ok, after several reboots/rescans it appeared correctly in my ESX Environment. Write operations to the new iSCSI Target (500GB) was also ok, but reading a lot of Data caused the Starwind Server to hang and no more I/O was possible. After this i have removed everything again and put it in my Testenvironment (I should have done that in the first place)

This is what came out after a 3 day testing and experimenting period:

Any Disk Bridge (and imagefile) iSCSI Targets, where the Disk itself is greater then 160GB, caused Starwind Server Hangs. (No Logfilereading anymore possible, also no Entries after reboot)

ESX Server Log reports ScsiPath Timeouts

The Starwind Server ends up on 100% CPU Load, (Hardware Interrupts), no more interaction possible.

As mentioned above, this only happens on Disks greater then 160GB, with 160GB or smaller Disks all is ok. Also creating an imagefile instead of a Bridge does not change anything.


Anyone has an idea or similar experience,

thanks and regards

Harry

[Robert (staff)]

Hello Harry,

Please e-mail directly to support@starwindsoftware.com and attach the log files from C:\Program files\StarWind software\Starwind\logs\ - this looks serious and needing our deep investigation.

Thank you for your cooperation in advance.

Rob.

[Mitschele]

Hi Rob,

thanks for the response. I dont think this is a Starwind issue. I just wanted to know if anyone had something similar. Anyway, my eastern holidays are messed up already and i keep analysing the problem. I will, by end of this week give a status here about what i have found out. The Starwind logfile does not contain anything since the Server itself hangs with 100% CPU load. A tool (Processexplorer) showed me 100% CPU Busy on Hardware Interrupts. The funny thing is that all disks are fine if i use them locally. I have used them with NTFS and copied large images back and forth without problems. But as soon as i use them as iSCSI targets i got problems with disks bigger then 160GB.

I am using a Silicon Image SATA Raid controller with 2-Ports and the latest drivers and firmware.

regards

Harry

[Robert (staff)]

Hello Harry,

We have seen similar issues before. They were related to the storage and RAID drivers. The drivers may cause hardware IO failures. Usually driver update fixes the problem.

If you could send us some detailed information such as logs and system information to our support e-mail address, we would be able to investigate this thoroughly.

Thanks,
Rob

[Mitschele]

Hi Rob,

i somewhat fixed the Problem. I also assume its a driver issue of the vendor. But, everything works if the disks are used locally, as i said.

My Solution to the problem is not nice but helps me for a while. I only use 1 Disk in Starwind as Bridge Device, then all is ok, regardless of size.

I will keep tracking this problem but since i am just evaluating VMware Virtual Infrastructure 4, this has to wait. Anyway, i got ESX4.0 running with Starwind iSCSI Targets and i am happy so far regarding Performance, etc....


Harry

[Robert (staff)]

Hello Harry,

We are glad the issue is resolved at least somewhat.

Please keep us updated if you have news.

Thanks
Rob