KernSafe Technology

Hello,

I'm seeing lots of scsi errors in the message log on my linux client while formatting my target filesystem (ext3), using a bridged device:

Jan 2 22:54:50 hercules kernel: end_request: I/O error, dev sdc, sector 10750031
Jan 2 22:54:50 hercules kernel: sd 33:0:0:0: SCSI error: return code = 0x08000002
Jan 2 22:54:50 hercules kernel: sdc: Current: sense key: Medium Error
Jan 2 22:54:50 hercules kernel: Add. Sense: Id CRC or ECC error
Jan 2 22:54:50 hercules kernel:
Jan 2 22:54:50 hercules kernel: end_request: I/O error, dev sdc, sector 10750287
Jan 2 22:54:50 hercules kernel: sd 33:0:0:0: SCSI error: return code = 0x08000002
Jan 2 22:54:50 hercules kernel: sdc: Current: sense key: Medium Error
Jan 2 22:54:50 hercules kernel: Add. Sense: Id CRC or ECC error
Jan 2 22:54:50 hercules kernel:
Jan 2 22:54:50 hercules kernel: end_request: I/O error, dev sdc, sector 10750543
Jan 2 22:54:50 hercules kernel: sd 33:0:0:0: SCSI error: return code = 0x08000002
Jan 2 22:54:50 hercules kernel: sdc: Current: sense key: Medium Error
Jan 2 22:54:50 hercules kernel: Add. Sense: Id CRC or ECC error
Jan 2 22:54:50 hercules kernel:
Jan 2 22:54:50 hercules kernel: end_request: I/O error, dev sdc, sector 10750799
Jan 2 22:54:50 hercules kernel: sd 33:0:0:0: SCSI error: return code = 0x08000002
Jan 2 22:54:50 hercules kernel: sdc: Current: sense key: Medium Error
Jan 2 22:54:50 hercules kernel: Add. Sense: Id CRC or ECC error
Jan 2 22:54:50 hercules kernel:
Jan 2 22:54:50 hercules kernel: end_request: I/O error, dev sdc, sector 10751055

Initiator is Linux CentOS 5.3, 64bit, running iscsi-initiator-utils 6.2.0.871-0.10.el5.

Target is Win2k8 r2, 64bit, running iStorage x64.1.58 trial. Target drive is raid5 array on a 3ware 9650se controller. all read/write caching disabled for that unit. It seems that I can partition the drive fine, setting type to 83 Linux, and I've tried a single partition as well as multiple parts. Formatting using mkfs.ext3 /dev/sdc1 runs really fast for about the first 5%, then slows way down and starts generating tons of those scsi errors. It will eventually complete after several hours, but yields a corrupted filesystem. I generally just cancel it once I start seeing all the errors. Partitioning and formatting the raid5 array locally works great for ntfs stuff.

Running on an image based target is fine, just way too slow. Network is simple dedicated gbit lan.

Any ideas?

Thanks!
Jon

We are now testing for this issue, please wait for 1-3 days.

We have reproducted the problem, we will resolve it in a few days.

We have just released a new version of iStorage Server for fix this problem, you may download it at:
http://www.kernsafe.com/download.aspx