LP1000 Breakdown

[Steen Okkels Nørby]

My Netserver LP1000r has crashed - failure: NVRAM/DISC Mismatch. Hardware failure ? According to HP: yes.
The 1M Raid controller was change, and a new driver installed. The server crashed again after 2 days..

Secondly the motherboard was changed, the server crashed again.

Now HP arrived with a new server, so that only the processors, disc' and Ram were "old". The server still crashed as the first time.

Now the disc's were replaced. The server crashed......

Finally the processor and ram were replacd. And then the server.... yes you're right - "crashed"

Software failure ? Hmmm... an image of a identical server was put on to the disc'. Then the server worked for 2 days, and failed with a disc error (nvram/disc mismatch)as the first time.

Any suggestion ?

[Sean T. Craig · Accepted solution]

Hi Steen,

Despite what anyone tells you, an NVRAM/DISK Mismatch doesn't automatically mean hardware failure. There are 2 types of mismatches: Normal and Unresolved. A normal mismatch means the configuration written to the NetRAID-1M card isn't the same as the configuration written to the scratchpad on the drives.
An Unresolved mismatch means there is a corrupted scratchpad on one of the drives.
The first thing we need to do is figure out what type of mismatch it is.

If it is a normal mismatch, you can re-synchronize them by going into the NetRAID Express Tools, choosing "Configure" --> "View/Add/Delete Configuration" then selecting each of NVRAM and DISK to verify which configuration is the correct one. If Disk is correct and NVRAM not, say "Yes" to save the configuration when exiting the DISK config and vice-versa if the NVRAM is correct. If you keep losing the configuration, I would be inclined to think it is something wrong on the controller side but as you've stated, that's already been replaced. Did they use the cache module off the old card when they replaced it or did it have it's own?

In the case of an "Unresolved Mismatch" we need to figure out which drive has the corrupted scratchpad. Go into the Express Tools and fail the first drive in the array. If you still get the mismatch, force that drive back online and fail the next one. Do this until you find which drive clears the error. Then you can rebuild that drive.

Hope this is helpful,

Sean T. Craig Sr. C.E.T., M.C.P.

[Steen Okkels Nørby]

Hi Sean,

We have changed the disc's. We did it in the following way.

Removed one disc, and let the server boot on the other, then I took another dics an rebuiled on that one. The server failed again.

Then I removed the one that hadn't been removed, and replaced that ande rebuiled on that. The server failed. As the last thing, I removed both disc, took a disc from a identical server (both sw and hw were the same), made a sysprep, and put into my domain. Finally when it was working as surposed, I rebuilded on a second disc. I'm sorry but the server failed again. The last disc was one of "old" one's, that had been used in the first rotations. Can the bad corrupted scratchpad be moved from one disc to another.

In the first cases it was the disc slot 2 that always failed, after I've taken the dics from the other server and put that in to disc slot 2, it's disc slot 1 that shows the "red light" and beebs... sector. Have I miseed something ? - during the setup ?