This post is more than 5 years old
4 Posts
0
36578
PE 2850 stuck at mptbios 5.06.06, not booting
Hello everyone
I came to work this morning and found our SQL server dead. On the screen it's stuck at the mptbios and it won't spin the devices. There's no raid setup on it, they're pure SCSI from the BIOS.
Can't even get into the bios, can't boot from the CD, nothing. I've noticed that even loading the bios and getting to the lsi mptbios is taking longer than usual. I might be wrong though. I've taken the drives out to see if I can get past that screen, with no luck , the mptbios is searching the luna's, can't find anything, because there aren't any drives in and it just stops. Does not let me into the bios, nothing.
I have currently 5 drives in it, 2x36gb and 3x72gb. One of the 36 is the os itself the others SQL db's and logs and 2 are empty I believe. I have however another pe2850 with the same lsi mptbios on it, I think the CPU is not the exact frequency but other than that they match. Should I just unplug the drives from this one and let it run in the new one while I figure this out?
I've looked everywhere for an answer couldn't find anything. It just died last night, I found it like that in the morning and there was no power outage or anything.
Any advice?
Thanks for reading my post.
theflash1932
7 Technologist
7 Technologist
•
16.3K Posts
1
February 16th, 2012 10:00
"Should I just unplug the drives from this one and let it run in the new one while I figure this out?"
You should be able to, if you like. IF YOU ARE SURE that RAID was not installed/enabled on the original machine, then you should be able to do this without a problem by setting the Embedded SCSI to SCSI Enabled.
Is your LCD panel amber and scrolling and error message(s), or is it blue?
You might try clearing the NVRAM using the jumpers on the motherboard near the system memory.
cout3
4 Posts
0
February 16th, 2012 10:00
Yes, I am sure there is no RAID on it. Funny thing, I had raid on it until 2 weeks ago and it worked just fine and I said to myself why complicate it further with this new install, let's keep it simple, I have some spare drives around, if one fails, restore from backup, blah ,blah,let's give up raid, and now this :) .
The panel is flashing all the time because of the power cable missing on one of the ps's, I use only one ps, in case it blows, I just switch the cable,I have no errors from the LSI board, it's not we're looking for the 5 nines of uptime here but still....
When you say NVRAM you mean the ram of raid board right ?
theflash1932
7 Technologist
7 Technologist
•
16.3K Posts
0
February 16th, 2012 10:00
If the LCD panel is flashing blue, hit the 'i' (Identify) button to turn of the flashing and see what the status/message is.
You know the purpose of dual power supplies is that when one fails, it automatically kicks over to the other one to keep the system up with no down time, right? (But it only works when their both plugged into power :)) Uptime might not be your utmost concern, but it could easily corrupt data in the crash.
NVRAM is Non Volatile RAM - it is where the BIOS settings and hardware lists are kept (the reason for a CMOS battery); when BIOS is not behaving properly, clearing the NVRAM will cause BIOS to rediscover and rebuild these lists on boot. The RAID controller has an NVRAM chip as well, but not jumpers; as you aren't using RAID anyway, we aren't talking about the NVRAM on the controller.
cout3
4 Posts
0
February 16th, 2012 11:00
Hey man, can't thank you enough. I cleared the bios on the MB and I unplugged the drives to see if it passes over the lsi screen, and success. From then on I was home free.
What happened was that this server was set to update automagically from ms and last night they must've pushed new packages in and it caused it to reboot.
I can't find however any reason why would the lsi board be stuck like that. I even kept it powered of for a few minutes as I thought a cold boot would do the trick.
Only resetting the jumpers helped. It was interesting because the LSI board would sometimes stop at lun2 and get past lun0,1. I don't know but I'm taking the rest of the day off as I nearly had a heart attack.
I imagined myself restoring and/or recreating the sql and attaching the db from backup, man what a nightmare.
Thanks man, I want to work where you work :) .
Thank you.
theflash1932
7 Technologist
7 Technologist
•
16.3K Posts
0
February 16th, 2012 12:00
The BIOS hardware and settings list became corrupted, so it probably got stuck trying to figure out what some hardware entry was ... which would make the MS updates merely a coincidence. The problem may have been lurking just waiting for a reboot ... or happened during the reboot.
Glad you got it back up and running :)
P.S. If you saw my paycheck, you wouldn't want to work where I do :)
cout3
4 Posts
0
February 16th, 2012 13:00
ROFL , oh man that was good!!!! Wait until you see mine :) .
Thank you again.
theflash1932
7 Technologist
7 Technologist
•
16.3K Posts
0
February 16th, 2012 13:00
No problem :)
jlms2011
2 Posts
0
January 8th, 2014 02:00
We have had a number of these servers do this on us. We contacted global services and they went round and round about how it was something we did. Well they finally did a warranty on one of the risers. This did fix the problem then it came back. This continued for a few times on different servers.
End of the day they told us to stop using hardware RAID to use it in scsi mode. Which BTW does work. Reset NVRAM turns off the controller. Go back and turn it on to RAID in any form or fasion and it comes right back. Again, something we are doing they say. Turn it on to scsi and it works. But what about the RAID. Hello ????? But sir in a busted broken accent you can do RAID from the OS................REALLY That is the solution. We did it now go fix it in your OS. Come on. If anyone knows anything more about this and has come up with a better option please let me know.
Until then we wont be spending anymore money with dell. See above and below.
Order a warranty drive for a Compellent SAN and they send us a DOA. Okay it happens. Send another one and it was DOA. Sent yet a third and sent it with the wrong customer name (going to a DC we rent Cabinets at, took site 2 hrs to find because of this). In the end three trips for the contractor we use for remote hands and wasted time all billed to us. Then Dell charged us for each drive. Two were returned and third was warranty (all actually). What a billing nightmare this was.
When we set up another compellent the Global services tech who came out used OM3 glass when we requested OM4. He also used 10 meter and left them everywhere in the back of the cabinet. Thus contractor replaced those and had to source them that day at full retail. Fine he used the 3 but you should have seen the mess he left. Very poor, a couple nice turns and a few pieces of hook and loop would have been nice.
Same compellent setup, sent us way more gbics than we requested or used. Enough to fill every unused port on the brocades. Sent us extra power supplies we didn't need, etc well you might need these they said. Round and round to return the stuff. T
Dell it is time to step it up, you are driving customers away. How about findning a solution to the problem this forum is listing here.
Again anyone else has a solution I would love to hear it. I would like to rrepair one of these 2850's and donate. Just trying to avoid buying another card out of my own pocket