February 26, 200917 yr I just finished my server build yesterday. I began by installing two drives without a parity, then moved my files from my old server to the two drives. I just installed the last drive, set it as Parity, and started the Parity sync. It seems to me the web control panel is very slow now that the parity sync is starting, and it says it's going to take, well basically FOREVER. Details: Total size: 976,762,552 KB Current position: 2,021,336 (0.2%) Estimated speed: 61 KB/sec Estimated finish: 262115.2 minutes Any help/advice?
February 26, 200917 yr I just finished my server build yesterday. I began by installing two drives without a parity, then moved my files from my old server to the two drives. I just installed the last drive, set it as Parity, and started the Parity sync. It seems to me the web control panel is very slow now that the parity sync is starting, and it says it's going to take, well basically FOREVER. Details: Total size: 976,762,552 KB Current position: 2,021,336 (0.2%) Estimated speed: 61 KB/sec Estimated finish: 262115.2 minutes Any help/advice? Going to need more info. Motherboard type, kind of drives, size of drives, if they are plugged into a card or motherboard, and anything else you think relevant.
February 26, 200917 yr Author Motherboard: Gigabyte GA-MA74GM-S2 (recommended on here and positive results via wiki and forum posts) Drives: Parity 1TB WDEADS Other: 2x 640GB WDAAKS Drives are plugged into the motherboard, it has 6 SATA ports. Like I said, the server was up and running fine without a parity not 30 min ago. I shut it down, put in the 1TB drive, plugged everything back in, started it, assigned the 1TB to parity, started the array, and it started the sync.
February 26, 200917 yr Author Apparently now I'm getting disk errors on one of the 640GB disks. I stopped the array, restarted the server, and ran parity sync again. It started fine and fast, but once these "disk" errors came up now it's slow and parity is taking forever. I don't understand how disk errors pop up when there weren't any errors 30 min ago when I transferred 600gb of files, and the disk didn't move. The only thing that changed was adding the 1TB parity. I may have changed cables of the disks around, but not the actual cables.
February 26, 200917 yr Apparently now I'm getting disk errors on one of the 640GB disks. I stopped the array, restarted the server, and ran parity sync again. It started fine and fast, but once these "disk" errors came up now it's slow and parity is taking forever. I don't understand how disk errors pop up when there weren't any errors 30 min ago when I transferred 600gb of files, and the disk didn't move. The only thing that changed was adding the 1TB parity. I may have changed cables of the disks around, but not the actual cables. If you could capture your syslog (directions here) and post it that would help. You will probably need to use a site like pastebin to upload the file.
February 26, 200917 yr Wow - that is really slow. Should be roughly 1000x faster! I would recommend taking smartctl reports on your drives and posting them as well. Follow the "troubleshooting" link in my sig for instructions.
February 26, 200917 yr Author Apparently the syslog is too large for pastebin. I've never heard of that, but it keeps giving me an error when I try to post it. As for the smart reports, I keep getting this error after telnet and trying to run smartctl: smartctl: error while loading shared libraries: libstdc++.so.6: cannot open shared object file: No such file or directory Sorry for all the trouble, just bothering me why this is going on.
February 26, 200917 yr If the syslog is too big, just post the first 200K of it or so. You likely are getting the same error repeating over and over after that. It is not necessary to see all of that detail. Often the telling information is available at the beginning of the syslog. There is a missing library from unRAID that smartctl need that causes the error you are seeing. Follow these instructions. (You do not need to do step 5, as unRAID already contains the most recent version of smartctl). http://lime-technology.com/forum/index.php?topic=2097.msg16658#msg16658 You can also use the package manager component of unmenu if you have it loaded.
February 26, 200917 yr Author Syslog: http://pastebin.com/m3a3c5b6 I stopped copying once it hit the repeating error. Smart reports: http://pastebin.com/m43a0c2fb http://pastebin.com/m2bc73b51 http://pastebin.com/mbd331ba
February 26, 200917 yr The good news is, your new parity drive and Disk 1 are fine, look good. But the SMART report for Disk 2 (serial ending 865), and corroborated by the syslog, looks very ominous. The 2 numbers that are particularly standing out are the Reallocated_Sector_Ct of 122 and the Current_Pending_Sector of 148. You can use the SMART report for the other 640GB drive as a baseline for comparison, as it is near perfect. Because this drive appears to have a large number of spare sectors in reserve, it is not close to total failure yet, but *may be* rapidly on the way to it. The syslog reports numerous media errors with the error flag of UNC (UNCorrectable). I would cancel the parity sync and un-assign the parity drive, because until it has completed a good parity sync, it is useless and will slow everything down. Then I would run a SMART long test on sdc, described in the Troubleshooting page, Obtaining a SMART report section. What we want to see on a subsequent SMART report is the Current_Pending_Sector RAW_VALUE go to zero, and STAY there, and the Reallocated_Sector_Ct STAY at a new value, with no additional increase at all. It WILL increase after the SMART long test, but we don't want to see any further increase after that. At this point, and based on what you see after this test, you may prefer to RMA the drive, and replace it with a new one. If you decide to continue with it, I would copy all files off it, and take the drive offline, un-assign it and run several passes of the Preclear Disk script on it. It will either kill the drive or pass it, and that will give you a good idea as to whether putting up with this drive is a good idea or not. If you then decide that the drive appears to have remapped ALL of the bad sectors, and is now usable, you will still want to monitor it for quite awhile.
February 26, 200917 yr Author I'm running the long smart test now. I'll post here and see what everyone else thinks, but most likely I'll RMA it. I just want to make sure I get the data back off it first, since the parity never completed yet.
February 28, 200917 yr Author I RMAed the drive. Bad: I'm impatient, kind of angry this happened as I'd just like my server up and running. Good: New drive, Western Digital has smooth RMA process, drive shipped same day. Will update if any more problems arise.
March 4, 200917 yr Author I received the RMA drive today. Now I've got another problem. I removed the 1TB drive from unRAID and set it up again as a USB drive to backup the two 640gb disks. Problem is now that my computer freezes when transferring data (assuming because the disk is freezing and has errors). Am I completely screwed out of my data now? I haven't even added the new 640gb drive because I was going to move the data back to the 1TB, swap the bad HD for the new one, retransfer the files back over and then set the 1TB as parity. Any advice would be greatly appreciated. I'm going to die if I lose any of this data.
March 4, 200917 yr Earlier, you said you had 6 SATA ports on the motherboard. Would it not be better to use them rather than a USB connection? For safety, you can disconnect the cables to your good Disk 1, or just make sure you don't assign it. You can keep the terabyte drive hooked up for temporary storage, but a better plan would be to hook up the new 640GB drive, run Preclear Disk on it so you know it's a good disk, then assign it as Disk 1 and click the Restore button, Start the array (click to format the new drive), and copy files directly from the bad Disk 2 to the new drive, temporarily installed as Disk 1. Once you have gotten all of the data copied over that you can, disconnect and reconnect the disks the way you want them, and reset the configuration again (click the Restore button). In other words, you are creating a temporary array for data recovery, then re-creating your array the way you want it. I would probably assign the terabyte parity drive last, in a separate step, to make sure the new array is good first, before you add the parity drive and start the parity build.
March 4, 200917 yr lishpy - BE VERY CAREFUL. When you are in a situation where you have unbacked up data you really need to think through each step that you are going to take. If your data is valuable (which I know it is), you may want to abandon the array temporarily and use your Windows workstation to backup your data. This may not be the quickest way, but it's likely what I would do. You need to install a ReiserFS driver in Windows. See this post. You should then be able to put your drive in a USB enclosure, hook it to the Windows machine, and see the contents. I WOULD NOT WRITE TO THE DISK!!! Copy the data from the disk. You might copy the files in chunks. If you get hangs or other problems, try to isolate the problem files and copy the others. The goal is to get as much data backed up as possible. Repeat this process for all of your disks that contain data. Now you will be able to go back and troubleshoot unRAID without fear of further data loss. In general it is a good idea to keep all of your data backed up. This is especially important while burning in a new server. Everyone that does not have their data backed up has to realize that there is a risk of losing that data. At its best, unRAID is not a backup solution. And until the server is burned in and stable, it is not a particularly safe place to store data. Please realize that the forum (besides Tom) is an all volunteer force. You need to take steps to protect your data from loss that might occur as we try our to give you the best advice we can.
March 5, 200917 yr Author Thank you everyone for the help, I greatly appreciate it. I recovered all of my data except 2 files, which luckily were marginally important and I can live without. I'm attempting to run the preclear script, but I'm not sure on exactly how to begin running it. Where do I move the script to? I downloaded the file and moved it onto my flash, telnetted to tower and am getting the error No such file or directory. It may either be in the wrong place, or I simply don't know what I'm doing wrong Again, I truly appreciate everyone's help. I switched to unRAID simply because of the strong community.
March 5, 200917 yr What is the command that you tried? And what is the device ID for the drive to be cleared?
March 5, 200917 yr Thank you everyone for the help, I greatly appreciate it. I recovered all of my data except 2 files, which luckily were marginally important and I can live without. I'm attempting to run the preclear script, but I'm not sure on exactly how to begin running it. Where do I move the script to? I downloaded the file and moved it onto my flash, telnetted to tower and am getting the error No such file or directory. It may either be in the wrong place, or I simply don't know what I'm doing wrong Again, I truly appreciate everyone's help. I switched to unRAID simply because of the strong community. This is not enough detail as to what you did and what the results were. You need to be much more methodical about what you did and what you observed. These kinds of things are hard enough to figure out with all the facts. Just getting 4 lines just isn't enough. Think of it as a 500 word essay. Give a status of each of the disks and if / where it is backed up. If there are errors on the disk, explain how you observed them (screenshots can be useful when you get nasty errors and want to get help understanding them). Retake smartctl reports of all of your drives. After that and before you do anything else, I'd suggest that we dialog and define a series of steps that you'll go through. Randomly running preclear scripts is likely to make things worse. BTW, If you put the flash disk in your Windows machine and downloaded a shell script to the root of the flash, when you boot unRAID it will appear in a subdirectory called "/boot".
March 5, 200917 yr Author What is the command that you tried? And what is the device ID for the drive to be cleared? The device ID is sda. I unhooked the rest of the disks in the server just in case to not lose any data. This is not enough detail as to what you did and what the results were. You need to be much more methodical about what you did and what you observed. These kinds of things are hard enough to figure out with all the facts. Just getting 4 lines just isn't enough. Think of it as a 500 word essay. Give a status of each of the disks and if / where it is backed up. If there are errors on the disk, explain how you observed them (screenshots can be useful when you get nasty errors and want to get help understanding them). Retake smartctl reports of all of your drives. After that and before you do anything else, I'd suggest that we dialog and define a series of steps that you'll go through. Randomly running preclear scripts is likely to make things worse. BTW, If you put the flash disk in your Windows machine and downloaded a shell script to the root of the flash, when you boot unRAID it will appear in a subdirectory called "/boot". What I did was moved each file, one by one, off disk2 (the 640gb disk with the SMART errors), and moved them on disk1 (the 640gb without any SMART errors). I didn't need to utilize any special methods in recovering the data. It just so happened that the first two files on the disk2 640gb HD were producing errors (as listed in my previously posted SMART reports). After trying the rest of the files, I successfully copied them to disk1. Now disk2 has been removed from my server, and sitting in the box waiting to be shipped back to Western Digital. I installed the drive that I received as my RMA replacement. This is the drive that I want to run preclear scripts on. Not randomly. Server situation: disk1: 640gb WD (with recovered data) [disconnected from the server] disk2: 640gb WD (RMA replacement) [connected to server, unformatted, waiting to run preclear scripts] disk3: 1TB WD (parity drive) [disconnected from the server] Like I said, I successfully copied my files off the bad 640gb disk, to the functional 640gb disk. That isn't a problem anymore. I disconnected the 640gb disk1, and the 1TB parity, and am now waiting on clarification on running Preclear scripts. I'm sorry but your response just threw me off and I'm not exactly sure if we're hitting the same page. Also I should clarify that I DID in fact move the preclear_disk.sh to the flash drive, but it did not move it to a folder named "boot." The plan is to preclear disk2, reconnect disk1, start the array, then assign the parity drive so I have parity.
March 5, 200917 yr Author Just to reference, I have the preclear script about to run on sda. My mistake: I didn't notice I hadn't changed the directory to "boot" At first, I had: root@Tower:~# ./preclear_disk.sh -c 1 /dev/sda Then changed to: root@Tower:/boot# ./preclear_disk.sh -c 1 /dev/sda
March 5, 200917 yr What I did was moved each file, one by one, off disk2 (the 640gb disk with the SMART errors), and moved them on disk1 (the 640gb without any SMART errors). I didn't need to utilize any special methods in recovering the data. It just so happened that the first two files on the disk2 640gb HD were producing errors (as listed in my previously posted SMART reports). After trying the rest of the files, I successfully copied them to disk1. Now disk2 has been removed from my server, and sitting in the box waiting to be shipped back to Western Digital. I installed the drive that I received as my RMA replacement. This is the drive that I want to run preclear scripts on. Not randomly. Server situation: disk1: 640gb WD (with recovered data) [disconnected from the server] disk2: 640gb WD (RMA replacement) [connected to server, unformatted, waiting to run preclear scripts] disk3: 1TB WD (parity drive) [disconnected from the server] Like I said, I successfully copied my files off the bad 640gb disk, to the functional 640gb disk. That isn't a problem anymore. I disconnected the 640gb disk1, and the 1TB parity, and am now waiting on clarification on running Preclear scripts. I'm sorry but your response just threw me off and I'm not exactly sure if we're hitting the same page. Also I should clarify that I DID in fact move the preclear_disk.sh to the flash drive, but it did not move it to a folder named "boot." The plan is to preclear disk2, reconnect disk1, start the array, then assign the parity drive so I have parity. [rant]I've been trying to help several people with problems and when the responses come back short and not very clear, it's diffiuclt to understand the current state which is frustrating. Also, you never have to follow suggestions made, but if someone takes the time to make a suggestion, you should at least acknowledge it and say you did or didn't take it and a short why. Just leaving it a big question mark is frustrating. For example, I still don't know whether you copied data to a Windows machine or not. I think not. Just say "I didn't have enough space on Windows" or "I feel confident that disk1 is solid and copied the data there." Saying that is much better than just ignoring the suggestion.[/rant] Sorry bout that. Feeling better now. So disk1 now contains all the data that was on disk1 + all the data that used to be on disk2. There are no backups. I'm assuming that you copied the data from unRAID to unRAID using MC or cp commands. Did you do any verifications with fc or md5 calculations to make sure that the files copied correctly? Is this basically the state? Your next step is to run preclear on the new disk2 that is currently unformatted. And then after that you will reassign the parity disk and build parity. And then after that you'll run a parity check. A backup would be better - but this sounds pretty reasonable. Just hope disk1 is solid for you. Good luck!
Archived
This topic is now archived and is closed to further replies.