February 22, 201016 yr Logged into my server and disk1 was red and had a few hundred errors.. Usually when a disk shows bad in a raid array I reseat the drive, reboot the server and see if the Raid rebuilds.. If it does, great, if it doesn't then I swap the drive and move on.... So with Unraid I decided to do the same thing, I did the restore option and now all the disk shows green but the parity drive is red... Its starting to show errors on disk1 yet again so I believe its safe to assume that this drive is now bad. My concern is the Parity drive now shows red and all the disk drives, including the one with errors shows green, if I remove disk1 and insert a replacement drive does the parity drive now think its bad and would refuse to rebuild? Or would it rebuild with the information it has stored? Note, after running the Restore option it initiated a parity rebuild which i canceled after a few minutes because it was going at like 8kb/sec. So am I screwed? Or should unraid be smart enough to rebuild disk1 with what information is on the Parity drive despite the parity drive currently showing Red...
February 22, 201016 yr Logged into my server and disk1 was red and had a few hundred errors.. Usually when a disk shows bad in a raid array I reseat the drive, reboot the server and see if the Raid rebuilds.. If it does, great, if it doesn't then I swap the drive and move on.... So with Unraid I decided to do the same thing, I did the restore option and now all the disk shows green but the parity drive is red... Its starting to show errors on disk1 yet again so I believe its safe to assume that this drive is now bad. My concern is the Parity drive now shows red and all the disk drives, including the one with errors shows green, if I remove disk1 and insert a replacement drive does the parity drive now think its bad and would refuse to rebuild? Or would it rebuild with the information it has stored? Note, after running the Restore option it initiated a parity rebuild which i canceled after a few minutes because it was going at like 8kb/sec. So am I screwed? Or should unraid be smart enough to rebuild disk1 with what information is on the Parity drive despite the parity drive currently showing Red... IF you pressed restore then you may have just fubar'ed parity. The restore button does not restore data at all; it sets a new disk configuration and throws out the parity that exists. It actually says something like the above next to the button. What you wanted to do was press the start button to bring the array online and recreated the contents of the drive that failed. The text beside each option would have said something similar. You may be able to get the contents off still but it may involve taking the drive out the array and trying to get to the files via another computer. Some of the more experiences and knowledgeable users here should be around soon to help.
February 22, 201016 yr Logged into my server and disk1 was red and had a few hundred errors.. Usually when a disk shows bad in a raid array I reseat the drive, reboot the server and see if the Raid rebuilds.. If it does, great, if it doesn't then I swap the drive and move on.... So with Unraid I decided to do the same thing, I did the restore option and now all the disk shows green but the parity drive is red... Pressing "restore" does NOT do what you think it does. It immediately invalidates parity and sets a new disk configuration based on the currently assigned and working disks. Basically, unless you immediately stop any parity calc that is under way, you just threw away any chance of recovering any data if the drive that failed is really failed (as opposed to a loose cable) Its starting to show errors on disk1 yet again so I believe its safe to assume that this drive is now bad. That is bad news, unless the power cable or SATA cable to the drive is loose and can be re-seated. Or unless the port on the disk controller is bad. My concern is the Parity drive now shows red and all the disk drives, including the one with errors shows green, if I remove disk1 and insert a replacement drive does the parity drive now think its bad and would refuse to rebuild?Parity is bad because you asked it to be invalidated by pressing the button labeled as "restore" It will not turn green again until you re-calculate parity (which would have happened if you subsequently pressed "Start") or taken a specific series of steps to make your array trust that your parity disk is OK. Or would it rebuild with the information it has stored? Note, after running the Restore option it initiated a parity rebuild which i canceled after a few minutes because it was going at like 8kb/sec.It is good you cancelled the parity calc, but it it still probably managed to overwrite the first few megabytes of parity data. The files in the equivalent sectors on the failed drive may be lost forever. Since the first sectors on any of the drives is the file-system itself, it probably is trashed the file-tree. So am I screwed? Or should unraid be smart enough to rebuild disk1 with what information is on the Parity drive despite the parity drive currently showing Red... unRAID would have been smart enough if you had used the "Start" button initially. It would reconstruct contents on an old drive to a new replacement. Now you are faced with a series of manual steps to try to recover. You may be lucky, but you may have lost files. Only way to know is to follow the steps below. 1. Capture and post a current syslog. Attach it to your next post. 2. Do NOT start the array, not until we better understand what you have done. 3. Do not ever press the button labeled "restore" unless you are removing a disk from the array, and not replacing it or changing the hardware to use the disks on never before used ports. (There is an exception to this rule, and we will need to use the button in combination with a special command to force whatever is left of your parity disk data to be trusted. 4. Do NOT add or delete any hard disks at this time. To perform ANY recovery, all must be present that were there when parity was last calculated. 5. Power down the server... Double check the power cable and SATA cable to the failed drive are securely attached. Un-plug and re-seat them. Then power up. It will not start. Do not start it. First we want to perform a few tests on the "failed" drive. Leave it in the array for now. 6. If you strongly suspect the drive is failed, purchase a new drive of equivalent size or larger, but not larger than your current parity drive. We'll need it later to try to reconstruct onto. Don't install it just yet. Our best hope is that the failure was just a loose cable or loose power cable, or loose power splitter. We'll check by running some SMART reports on the failed drive. If it is OK, we'll get it back online. If it is NOT, you will need to install a replacement drive and we'll have to force the unRAID server to think parity is GOOD, even though you overwrite the leading part of it. Then, we'll need to run file-system check/repair utilities on the re-constructed disk to rebuild the file-system damage. Only then will you know how much you lost by overwritting parity. Joe L.
February 22, 201016 yr Author Yeah I was kind of leaning on the side of screwed myself.. The restore option said the Data would be untouched which is why I used that option... As I mentioned I am used to normal Raid Systems, so to hit 'Rebuild' on a Raid5 drive would not kill anything if the data drive was proven to be bad on only the one disk... When I started the array it still showed red, which is why I went with the Restore option option, because rebuild/verify was not an option
February 22, 201016 yr When I started the array it still showed red, which is why I went with the Restore option option, because rebuild/verify was not an option A disk that has been disabled because a WRITE to it failed will show as red until you un-assign it on the devices page, start the array with it un-assigned (this causes the array to forget its model/serial number), stop the array once more, then re-assign it. Otherwise it thinks the same disk is still that that failed and will not put it back online. Instructions for posting a syslog are in the wiki under troubleshooting. You did not describe which version of unRAID you are running? Do you have any add-ons installed? As I said, until I can see a syslog it is harder to guide you. Just don't press the button labeled "restore" It is true, data on working disks is not touched, but data on failed disks is probably going to be lost as you re-calculate parity without it. The only hope is that your subsequent parity calc that was going at a very slow rate did not clobber anything important and that you stopped it in time. Joe L. Joe L.
Archived
This topic is now archived and is closed to further replies.