June 24, 201016 yr In your first post you said you moved your existing 9 drives into this 'new' case - did you have these 9 drives working fine in a different case? What is different besides the case? What is the exact model number of your power supply?
June 24, 201016 yr Author The system has been running in a tower case for over a year, with an occasional HD addition or upgrade. The 9 drives that were moved have been running steady for a few months, with no problem at all. Now, I moved to 4220 case with a new motherboard, that can provide multiple PCI-E slots to run SAS cards. The power supply is OCZ 700MXSP.
June 24, 201016 yr Author I have just moved 8 drives that were controlled by the SAS card to completely different row in the 4220 case. Disk one has a [glow=red,2,300]red[/glow] ball next to it and all other drives are [glow=green,2,300]green[/glow]. Syslog attached. syslog-2010-06-23_e.zip
June 24, 201016 yr Author I have performed initconfig followed by mdcmd set invalidslot 99, and now the system is running a Parity-Check. I guess I'll let it run through the process and report. The syslog in the prior post is the latest. I am exhausted... :'( :'( :'( . . . About 2% into Parity-Check the Disk 1 is red ball again and there are now 576 errors reported. There is now no option to run parity check... How do I make this problem go away??? This is beyond frustrating... If there is only one disk that is giving me problems now, can I move the data off of that disk onto the other drives and reinitialize that drive (disk 1) as new?
June 24, 201016 yr It almost sounds like a power supply issue, where having more drives spun up triggers the situation.
June 24, 201016 yr Author It almost sounds like a power supply issue, where having more drives spun up triggers the situation. I have another power supply that I can try... Besides, the last case had an old Ultra 500W PSU running without a single problem. I've performed 4 attempts at running parity-check and it is always aborting at around 2%, with 576 errors on disk 1. Seeing that all the errors seem to be attributable to disk 1, can I do anything to "fix" that drive? It seems that there are a few posts here and there that people are able to reconstruct the data from far worse.
June 24, 201016 yr Assuming the drive is actually bad, as long as you've been careful and Parity is still intact, you will be able to reconstruct it's contents. As mentioned by BRiT - this sounds like a power supply problem. In looking at that supply, it appears to be a 2-rail design (you can tell that by seeing on the label that +12V1 and +12V2 are both rated at 25A). This means the one of the rails is dedicated to the motherboard, and the other is dedicated to the "peripherals", ie, the hard drives. So even though you might think there is 50A available for your drives, actually there's only 25A. Now 25A should be sufficient for 9 drives, but if the PSU is at all marginal, then this could account for the problems. If in your shoes, next thing I would do is swap out the PSU with your old one. Actually, what I would do is probably set up your old m/b & psu on a bench & hook up all the drives & verify everything is ok. Then I would set up the new electronics with old psu on a bench. This is something for all system builders: You should always set up your PSU, electronics, and drives on a bench and test before assembling into a chassis. If something is not working, as is the case here, it can be very frustrating to troubleshoot, as well as a lot of work. On the bright side, I bet you know a lot more about how the server works now
June 25, 201016 yr Author Well... I have a little more of an update. It was getting to me this whole working-in-one-case-but-not-the-other thing, so... After the 4th time of parity-check aborting less than 2% into the scan I noticed something that I did not pay attention to. The temperatures of the HDs were all in the high 40's! I have never seen such high temperatures before and was puzzled by this. As I mentioned, I am housing my system in a new Norco 4220 case. I never gave it another thought when putting together the computer, but... I even replaced the stock 4 x 80 mm fans with 3 x 120 mm fans. When looking at the system with the top open I realized what the problem was as far as the high temps. The drives were not being adequately cooled because the air was not getting through the hard drive cages, but flowing from the top of the case into the fans, bypassing the drives! I closed the lid and ran parity check once more. 44 errors were found but the check completed in 12 hours and now I am running another one just for the heck of it. I did learn a few things. Go figure, high temperature was causing the problems? I cannot seem to attribute it to anything else as I did not change anything else at that point. I've attached the latest syslog. Thoughts? syslog-2010-06-24_b.zip
June 27, 201016 yr Author Just a follow up - I've been using unRaid now for the last several days, running parity checks and reading/writing to the array. So far, I have not had any issues. Again, thanks to everyone that helped. Lev
Archived
This topic is now archived and is closed to further replies.