DL140s (G1) corrupting NTFS drives

[Andy Broyles]

I have 10 DL140s 2x2.4Ghz and 2GB RAM (all original equipment) in a rack and all are running Windows 2003 std server SP1, Symantec Corp A/V 10.1, and Veritas Backup Exec 10.

The machines also each have an Adaptec FC1511 iSCSI card which connects them to an iSCSI SAN.

On three of the servers I am getting consistent NTFS corruption errors on their local Seagate ST380011A 80GB hard drives.

I have replaced the hard drives and the IDE cables on these machines and yet the problems still exist. Only the software listed above is common on all three machines...and I have uninstalled the Symantec and Veritas software to troubleshoot to no avail. The machines are running at a low CPU/memory usage level (occasional spikes of course) and nothing really consistent about when the corruption starts to show up.

I have also reinstalled the removed drives in other machines and have not experienced any of the corruption issues anywhere else.

I am thinking that the problem has to lie with the motherboard, but the 10 machines are sequentially serial numbered and the 'problem' machines are interspersed through out the series.

Thankfully, running a CHKDSK clears the problems pretty quickly, yet within 48 hours the errors start reoccuring. The corrected errors are consistently $I30 index type errors, which I believe deal with the hard drive's directory index bitmap...but I am not positive of that.

One other item of note...since the machines are connected to the SAN, the only thing installed on the local drives is the operating system, the temp directory and the pagefile. Disabling the Windows Write caching seems to help delay the problem, but it doesn't solve it. Interestingly, the servers still continue to run without problem in this configuration since they are primarily accessing their data from the SAN.

Any advice on how to further troubleshoot this will be appreciated...