April 29, 201115 yr I started a large copy last night, and came in this morning to a copy error. I opened up the unraid gui in my browswer, and there is 1 red-balled drive with errors, and 1 gree-ball drive with errors, both reporting 0 degrees temperature. I have run the smart reports, they are attached. Fortunately I have not copied anything important to the drives in weeks, just media that I can replace. What do I do next!? smart_Drive_1.txt smart_Drive_2.txt
April 29, 201115 yr Hello, Sorry I can't help you much as I am new, but I wonder if you could share a little more info: How long have you server been running How old are those drives? What brand and size are those drives What action have you taken sofar? How large is the file that you copy. These info will help me build up a feel for running the server. Thanks
April 29, 201115 yr Author Server has been running for many years, the drives that have errors are 1.5T Seagates, with current firmware (no upgrade was required I checked) and they are maybe 1-1.5 years old. I have taken no action, I saw the copy error which cancelled/stopped the file copy, then I ran the smart reports. Last parity check was over a month ago. I was doing a copy of about 150gigs of multiple files (blu-ray rips).
April 29, 201115 yr The SMART reports didn't work. You could try a power cycle and check the cables while the server is off. Maybe it's a common power splitter acting up. Otherwise, you'll have to try and figure out if the drives are alive or not. Maybe try unassigning them and then mounting them in unMENU if they will appear. I can't recall if unMENU will share a drive while the array is stopped. Peter
April 29, 201115 yr Author So what does it mean that the smart reports failed?! It is not a common power cable, my case is open and I can see they are on different power connectors. What to do next?
April 29, 201115 yr Author Also, what does it mean that only 1 drive is red-balled, the other is green-balled but 0 degrees celcius. When I try to access the individual drives, I still have access to the red-balled drive, but no access to the green-balled drive. I have purchased 2 new drives, but what are my next steps? Can I re-build one or both drives? Should I force unRaid to accept the drives? Run more tests? Why is smart drive not able to run on either of the drives?
April 29, 201115 yr Also, what does it mean that only 1 drive is red-balled, the other is green-balled but 0 degrees celcius. When I try to access the individual drives, I still have access to the red-balled drive, but no access to the green-balled drive. I have purchased 2 new drives, but what are my next steps? Can I re-build one or both drives? Should I force unRaid to accept the drives? Run more tests? Why is smart drive not able to run on either of the drives? Do not start pulling drives.... you'll only get into trouble. Red balled indicates a write to the drive failed so it was disabled. It may be perfectly fine. Right now its contents are being simulated by all the other drives in the array in combination with the parity drive. The "green" drive with errors are "read" errors. You can ignore those for right now. If you pull the green drive you'll lose the ability to simulate the "red" drive. Power down. Check and re-seat all the connections to the two drives, both power AND data. Then reboot and post a system log. DO NOT format any disks, or replace any disks until you get expert advice. The "red" drive will not go back into service on its own. It must be re-constructed. (remember, a write to it failed, so it is guaranteed to be out of date) Joe L.
April 29, 201115 yr Author Joe, I was hoping you would respond. I powered down the server, swapped out the power and data connectors off the 2 bad drives with known good drives, then booted it back up. There is just the 1 red-balled drive, the drive that only had "read" errors is still green, and now reporting its temperature. I then copied some critical data off, just in case. This is such good news, hopefully I can pull the red-balled drive and put in a new one and re-build, but will await instruction. Then I need to know what to do with the drive that had read errors. I have attached the syslog, ran it after I copied some critical files off, hopefully that does not screw you up. Thank you Joe! Rory syslog-2011-04-29.txt
April 30, 201115 yr Joe, I was hoping you would respond. I powered down the server, swapped out the power and data connectors off the 2 bad drives with known good drives, then booted it back up. There is just the 1 red-balled drive, the drive that only had "read" errors is still green, and now reporting its temperature. I then copied some critical data off, just in case. This is such good news, hopefully I can pull the red-balled drive and put in a new one and re-build, but will await instruction. Then I need to know what to do with the drive that had read errors. I have attached the syslog, ran it after I copied some critical files off, hopefully that does not screw you up. Thank you Joe! Rory The syslog does not show any errors on any disk (or at least i did not see them) Can you get a smart report on the failed disk? (it might have just been a loose cable that you've subsequently fixed.) If it responds to a smart report, you can try to reconstruct onto itself. To do that you must get the array to forget its morel/serial number. You do that by starting the array with the disk un-assigned. Stop the array un-assign the disk that is "red" Start the array with it un-assigned. Stop the array Re-assign the disk. Start the array . This time it will re-construct the disk back onto itself. When fully re-constructed, you'll be up as normal.
April 30, 201115 yr Author Thanks again for the help, the smart drive was able to complete, it is attached. Should I proceed with the re-construct with the instructions you posted? Thank you! smartSDD.txt
April 30, 201115 yr The drive shows defects: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 17 This value (17) needs to be monitored. If it increases then the drive is surely failing. I would proceed with the rebuild and then check this value. You'll need to monitor the Reallocated_Sector_Ct for a while. If it holds steady then the disk is OK. If it increases then the disk is bad.
April 30, 201115 yr Should I proceed with the re-construct with the instructions you posted? Also get a smart report of the "green" drive with the errors. If it looks reasonable, then you can proceed with the re-construction.
April 30, 201115 yr Author Here is the smart report for the drive that had the read errors on it. smartSDE.txt
April 30, 201115 yr It had one sector that was already re-allocated. you can start the process of reconstructing the drive that was disabled.
April 30, 201115 yr Author Sweet, thank you for the help Joe, greatly appreciated. Re-construction has started.
April 30, 201115 yr Hi, I really like to know if you were able to completely recover ?. is the array back online and operational ? how is the reconstruction progress going ? were the drive clicking while reconstruct? it is pretty cool to see expert in action. quick question how can I execute a smart report ?
May 1, 201115 yr Author Yes, with Joe's help I now have a fully functioning array again, without replacing the drive. Parity check completed without error, all drives are now green. No clicking during reconstruct that I heard, the drives are loud though, and got to 38 degrees, a 12 hour workout seems to do that! Smart report instructions are in the wiki, http://www.lime-technology.com/wiki/index.php?title=Troubleshooting#Obtaining_a_SMART_report Thanks again Joe!
Archived
This topic is now archived and is closed to further replies.