Win2k sees LUN twice on Autoraid 12H?

[Ralph Hearne]

I?ve got a question. Why does Windows2000 see a LUN on a 12H disk Array twice? Should it not see just 1 disk volume?

We have a Netserver LH3 connected to a 12H AutoRaid array via 2 Host Bus Adapters (HBAs). When I installed it 2 years ago, the NT4 operating system saw the single LUN on the array as a single drive letter (i.e. it was the D: drive.). Last year, while installing new 36G disks, an HP engineer updated the BIOS on the server and the Array firmware (to HP62) while he was here. After he left, I noticed that the NT4 disk manager began to see the LUN on the array twice ? i.e. as the D: drive AND again as the E: drive. Operations continued however as normal and we just ignored the E: drive since any changes made to one drive was reflected exactly on the other.

In April 2001, we upgraded the LH3 server to Windows 2000 ? what a disaster that was! Every time the server booted up, Windows 2000 ?plug-and-pray? saw NEW hardware (the LUN on the array), rebooted, and changed the security descriptors on all our files. Data was lost and eventually the array had to be reformatted.

Now I?m trying to get it ready for service again. I updated the LH3 BIOS (4.06.43 PL) and installed Windows2000/SP2 from scratch. It was stable. Then I powered on the 12H Array and cold-started the server. Although the PNP problem is fixed, the server disk manager still sees 2 LUNs. I (ntfs) formatted the D: drive (disk0) and (after a while) the E: drive (disk1) appeared to be formatted as well. Yet when I copy files to the D: drive ? explorer does NOT see them on the E: drive. Can anybody tell me why? Why are there 2 drives in explorer for 1 LUN in the array? What am I missing here?

Any advice would be very much appreciated.

Sincere Thanks,
Ralph Hearne
Network Administrator
Indal Technologies Inc.
3570 Hawkestone Road
Mississauga, Ontario
Canada L5C 2V8
Phone: 905-275-0012 ext.173
Fax: 905-273-7004
E-mail: ralph_hearne@indaltech.com

[Bill McNAMARA_1 · Accepted solution]

The autoraid on nt must use the autopath config.
I'm not sure of it in windows, but on unix the configuration for the array to prepare it for nt use is arraymgr -o

What this does is disable (no green light) the second controller and only activate it when the primary fails.... with the same scsi id as the one that has failed since nt can't switch like unix lvm.

In you case the autoraid controlers are more than likely on the same bus and both active.
One of them needs to be a passive failover.

Bill