February 22, 201214 yr I've had this happen about 5 times now, and I had posted a little bit of the problem in my unrelated thread ( http://lime-technology.com/forum/index.php?topic=18523.0 ) but I figured I would get more attention with a new thread by itself. When transferring files from disk2 to disk3 (disk2 has pending sectors and I need to preclear it then possibly RMA it), the server appears to hang and the drives become inaccessible. This happens after 10-20 minutes into the copy from disk2 to disk3 (~5GB transfered of 199). When the hang occurs, I am not able to access the tower through the web GUI or through telnet. When I plugged in a monitor and keyboard directly, the system was hung and not responsive to keyboard inputs. I took a screenshot of the last printed lines: Am I possibly trying to read a bad sector which is causing this? I've wanted to try and transfer the files via Midnight Commander, but can't figure out how to use it! Heh, see my thread on that here: http://lime-technology.com/forum/index.php?topic=18623.0 Any help would be great!!
February 22, 201214 yr Get another drive and rebuild from parity. Sent from my SGH-I727R using Tapatalk
February 22, 201214 yr Author Get another drive and rebuild from parity. Currently I have disks 1, 2, and 3. Can I unmount disk 3, pull out disk 2 and place disk 3 in it's spot and unRAID will rebuild it perfectly?
February 22, 201214 yr If you are sure ALL your other drives are good, then you can disconnect the one bad drive, and start the array with the disk being simulated. Then you can copy the data to another physical disk. Be VERY sure which physical drive you are disconnecting. Jonathan
February 22, 201214 yr Get another drive and rebuild from parity. Currently I have disks 1, 2, and 3. Can I unmount disk 3, pull out disk 2 and place disk 3 in it's spot and unRAID will rebuild it perfectly? No. Unraid only tolerates one bad drive. If you unmount disk 3, you will no longer have the complete set of information available to simulate or rebuild disk 2. You need another physical drive the same size as the drive you are replacing or larger up to the size of the parity drive to rebuild to.
February 23, 201214 yr Author Get another drive and rebuild from parity. Currently I have disks 1, 2, and 3. Can I unmount disk 3, pull out disk 2 and place disk 3 in it's spot and unRAID will rebuild it perfectly? No. Unraid only tolerates one bad drive. If you unmount disk 3, you will no longer have the complete set of information available to simulate or rebuild disk 2. You need another physical drive the same size as the drive you are replacing or larger up to the size of the parity drive to rebuild to. Sorry if I wasn't clear. I currently have 3x 2TB. One is parity, the other two are disk1 and disk2. I have a 1.5TB disk that is currently in the array, but has no data on it. Can I remove this 1.5TB disk from the array, pull out the bad 2TB disk (it only has 200GB on it), and replace the bad disk with the 1.5TB? Or, does the replacement drive need to be 2TB? (even though it isn't full) Otherwise, how can I get the data off of the bad disk onto a good disk?
February 23, 201214 yr Get another drive and rebuild from parity. Currently I have disks 1, 2, and 3. Can I unmount disk 3, pull out disk 2 and place disk 3 in it's spot and unRAID will rebuild it perfectly? No. Unraid only tolerates one bad drive. If you unmount disk 3, you will no longer have the complete set of information available to simulate or rebuild disk 2. You need another physical drive the same size as the drive you are replacing or larger up to the size of the parity drive to rebuild to. Sorry if I wasn't clear. I currently have 3x 2TB. One is parity, the other two are disk1 and disk2. I have a 1.5TB disk that is currently in the array, but has no data on it. Can I remove this 1.5TB disk from the array, pull out the bad 2TB disk (it only has 200GB on it), and replace the bad disk with the 1.5TB? Or, does the replacement drive need to be 2TB? (even though it isn't full) Otherwise, how can I get the data off of the bad disk onto a good disk? Answer... no... becaue once you take the 1.5TB drive out of the array you have two missing disks. Simple solution. Just copy from the "simulated" failed drive to the 1.5TB drive, and once your files are on it you can decide how to proceed. (Yes, a replacement for any existing drive must be as big as the existing drive)
February 23, 201214 yr These questions have already been answered. Can I remove this 1.5TB disk from the array, pull out the bad 2TB disk (it only has 200GB on it), and replace the bad disk with the 1.5TB? No: No. Unraid only tolerates one bad drive. If you unmount disk 3, you will no longer have the complete set of information available to simulate or rebuild disk 2. Or, does the replacement drive need to be 2TB? (even though it isn't full) Yes: You need another physical drive the same size as the drive you are replacing or larger up to the size of the parity drive to rebuild to. Otherwise, how can I get the data off of the bad disk onto a good disk? If you are sure ALL your other drives are good, then you can disconnect the one bad drive, and start the array with the disk being simulated. Then you can copy the data to another physical disk. Be VERY sure which physical drive you are disconnecting. Jonathan
February 23, 201214 yr Author Simple solution. Just copy from the "simulated" failed drive to the 1.5TB drive, and once your files are on it you can decide how to proceed. That's the whole point of this post - I'm not able to copy from the bad drive. What do you mean simulated? Thanks everyone for your help!
February 23, 201214 yr If you are sure ALL your other drives are good, then you can disconnect the one bad drive, and start the array with the disk being simulated. Then you can copy the data to another physical disk. Be VERY sure which physical drive you are disconnecting. Jonathan I didn't feel like retyping all of it. DISCONNECT THE ONE BAD DRIVE.
February 23, 201214 yr Author Ok, I disconnected disk2 and then began copying the files from "disk2" to disk3. It worked for the first 10 minutes or so and then the same problem/hang/freeze occurred! Windows said the server is unavailable during the copy, telnet doesn't connect, and the web GUI won't load. Only option was to power cycle the box.
February 23, 201214 yr Author Brought the server back up with disk2 inserted and now it is showing as "disabled" -- why's that? Sorry for all the n00b questions, as you can tell I'm learning all of this.
February 24, 201214 yr It's listed as disabled because the array was started without it. I assumed from what you posted that you were sure that you had one bad drive. Apparently you have other bad drives, and / or another issue locking the machine. Since you didn't /can't post a syslog, it's hard to know what's going on. I'd start with running a memory test for several passes. Posting smart reports for ALL the drives wouldn't hurt either. There was a reason I started my previous post with "If you are sure ALL your other drives are good". If you try copying from the simulated disk2 to disk1, does it still lock up? Jonathan
February 24, 201214 yr Author I'm fairly certain that the root of all these problems was a networking issue, but it's too late now to go back to the state I was in... or can I undo all of this? I was able to get the files off of the simulated disk2 and onto disk3, so I stopped the array and unassigned disk2 as I want to preclear it to see if it's truly a bad drive. Now it is showing the drive as "HPA?" plus it's listed below as an unmounted disk:
February 24, 201214 yr At this point unless someone else says otherwise, I think it's time to reinitialize the array without the missing disk. If you get another disk failure right now, parity will not protect you, as it is currently simulating disk2. If you reinitialize the array so it recalculates parity from disk1 and disk3, then at least you will be protected from another failure. If parity will not recalculate correctly now, you stand a good chance of losing data. If it were me, I'd dd zeros to the simulated disk2, which will update parity to a state where you can remove the disk assignment and still keep parity protection, but you will be at jeopardy of data loss the entire time you are writing zeros, and if you mess up the command to write zeros by 1 character, you can permanently erase the data you are trying to protect. Before you do anything else, I think I'd make sure you have backups of all important data copied off the array machine totally. You can't trust your unraid right now until you get the crashes sorted out.
February 24, 201214 yr Author Yes, I'd like to reinitialize the array without disk2 - how do I do that?
February 24, 201214 yr Author Yes, I'd like to reinitialize the array without disk2 - how do I do that? Ok, I found another thread in which initconfig was explained. I have run that and am rebuilding parity right now. Once the parity is rebuilt, I'll run a parity check to ensure there are zero errors. After that, I will run preclear on disk2 and see if the drive is bad. Thanks everyone, especially Jonathan.
Archived
This topic is now archived and is closed to further replies.