Netservers

NetRAID 1M controller losing NVRAM configuration...

 
SOLVED
Go to solution
KWenners
Occasional Advisor

NetRAID 1M controller losing NVRAM configuration...

We have a Netserver LP2000R with a NetRAID 1M controller and four 18.2GB drives (3 on a RAID-5 array, 1 hot-spare).

The problem is the server will frequently lock up with various "disk error" messages (WinServer 2003). During the reboot a POST message from the NetRAID controller reports an "invalid NVRAM configuration". Pressing CTRL-M brings us into the setup for the NetRAID controller. We select "View/Change Configuration" and the drive array configuration is listed with no errors (three drives "ONLIN", one drive "READY"). Press ESC, select "Yes" to "Save Configuration?" and reboot. The server works again as if nothing was wrong. But several hours or minutes later it will fail again.

We've reseated all of the cards, cables and memory. We verified that all of the fans are spinning and unobstructed. We loaded the latest BIOS on the NetRAID controller. The server is located in a raised-floor environment with plenty of cooling and power conditioning.

Is this a 'common' problem? It seems to be the NetRAID controller at fault, but we want to verify this diagnosis rather than shot-gunning parts. Any and all help greatly appreciated.
4 REPLIES 4
Bryan Eley
Trusted Contributor

Re: NetRAID 1M controller losing NVRAM configuration...

This is probably a real shot in the dark, but does your system have a Toptools card in it? There was an PC magazine article I read a while back regarding the LP2000r that talked about a toptools firmware issue causing raid controller problems.

The portion of the article of interest was:

"During our tests, the Netserver was plagued by mysterious false positive alerts, incorrect management data, and even the occasional loss of service. At one point, it stopped recognizing the RAID controller, which made loading the OS impossible. In another instance, Instant Toptools could not successfully measure temperature or voltage readings. According to HP, these problems were caused by a faulty revision of the firmware on the Toptools Remote Control card. During the POST process (initial boot), the card would occasionally grab all available resources, and the SCSI and RAID BIOS could not load. HP knows about these problems and expects to release a Toptools firmware fix on its Web site in the near future."

http://www.findarticles.com/p/articles/mi_zdpcm/is_200205/ai_ziff25893

I know this is a real stretch, that you may not even have a toptools card and that the problem quoted wasn't necessarily an exact duplicate of the problem you had, but thought it worth bringing up for such a (presumably) unusual occurrence.

Best,
Bryan
KWenners
Occasional Advisor

Re: NetRAID 1M controller losing NVRAM configuration...

Thanks Bryan, but we don't have a TopTools card installed in this NetServer. The server has worked for many years before this trouble popped up.
Andrew Pywell
Advisor
Solution

Re: NetRAID 1M controller losing NVRAM configuration...

The RAID configuration is saved to both the controllers NVRAM and each disk (COD). You appear to be recovering the configuration from the COD which implies a configuration mismatch message (as compared to an unresolved configuration mismatch). If this is the case, you should review (but not save) the NVRAM configuration choice on next failure. Check for drives with FAILED status rather than ONLINE. If they show as failed then the SCSI connection between the NetRAID and the SCSI bakplane and drives is intermittent, with a bad SCSI cable most likely for that system model.
KWenners
Occasional Advisor

Re: NetRAID 1M controller losing NVRAM configuration...

Andrew's reply seems to be the winner. We don't have a replacement SCSI cable so we simply removed and reinstalled the existing cable. Before, the server would fail after a few hours. Now, it's gone the whole weekend with no sign of trouble. Thank you.