July 10, 200817 yr I've been using unraid long enough now that one of my data drives has finally failed. It was a 250GB sata drive - actually all the drives in the array were identical (size and model). The smallest spare drive I have is a 300GB pata and a 1TB sata. I opted to use the 300GB pata and I do not seem to be able to get any "parity-swap" option to come up under unraid 4.2.4 pro. Perhaps I do not understand the instructions in the forum or the documentation. I checked AVS forum as well. Now I'm thinking that I should have started using the sata drive first and that I may have confused unraid into forgetting about my original 250GB failed drive. I started thinking that "swapping" the drives in the forum instructions meant "use the 'devices' page of the web utility", but it sounds more like a hardware swap in the "official" unraid manual. If it is a "hardware swap", then you can't very easily swap a pata and a sata and expect the device to look the same without some unraid hacking wizardry by an experienced linux sysadmin. So, if I use the "devices page" and swap the "parity" disk with the "disk2" disk and go back to the "main page", I do not get any special options. It just says "too many missing/incorrect disks". Here's where things stand now: Disk status Model/Serial No. Temperature Size Free Reads Writes Errors [Green] parity Maxtor 7Y250M0/Y63MGPEE 39°C 245,117,344 - - - - [Green] disk1 Maxtor 6Y250M0/Y696DF1E 36°C 245,117,344 - - - - [blue] disk2 ST3300620A/9QF2QPW8 31°C 293,036,152 - - - - [Green] disk3 Maxtor 7Y250M0/Y63C1R6E 37°C 245,117,344 - - - - [Green] disk4 Maxtor 7Y250M0/Y63MGL0E 34°C 245,117,344 - - - - [Green] disk5 Maxtor 7Y250M0/Y6300GSE 34°C 245,117,344 - - - - [black]Command area Stopped. Disk in parity slot is not biggest. If this is a new array, move the largest disk into the parity slot. If you are adding a new disk or replacing a disabled disk, try Parity-Swap. Restore will initialize the stored array configuration; all drives will appear as New, but data disk contents are not affected. I'm sure I want to do this
July 10, 200817 yr If you have a failed disk you DO NOT want to use the restore button. If you do, any ability to recover its contents will be lost. You are correct in that the official manual is out of date. It was written before the devices page existed, when the only way to swap device assignments was to physically move the cables. Clearly, you cannot when one is an SATA, and the other IDE. Instead, you need to swap their logical device assignments. I am assuming that disk2 failed on you. You want to use the 300Gig drive to replace it, but that is a PATA, while the original was an SATA. To do this, with the array stopped, assign the 300Gig drive as parity, and the old parity drive as disk2. Then, when you go back to the main page, you should have the option to do the parity swap. Whatever you do DO NOT PRESS THE RESTORE button. If you do, it will immediately delete the super.dat file (actually, it renames it, but to unRAID it is gone) and then rebuilds it based on the currently assigned drives, throws away any existing parity, and proceeds to calculate it new. The effect will be the loss of any data that was on the failed disk2. The "restore" button does not restore any data and in fact throws away any data in failed drives. Only data in assigned working drives is carried forward in the new configuration. Let the parity swap process complete (it will take a while, as first the old parity drive is copied to the new, and then the old disk2 reconstructed. It will probably be 3 or 4 hours or more, even with your faster SATA drives.) After parity calculation is complete, we can talk more about the 1TB drive. All you can use it for is parity (since it is bigger than all your other drives) If you have a spare slot on the SATA controller, you can connect it there, then assign it to parity (un-assigning the existing 300Gig IDE drive) and let it complete the parity calc process. At that point, you can assign the 300Gig drive as a data drive. (If you have the Pro license of unRAID and can handle more than 6 devices) Joe L.
July 10, 200817 yr One more thing... What version unRAID are you using? (you posted in the 4.2 thread) There was a bug in the parity-swap processing in version 4.2.1, and probably in earlier versions too. See excerpt from release notes below: I know this is not normally something you would want to do when you have a failed drive, but you really should upgrade to the current 4.3.2 release. Joe L. Changes from 4.2.1 to 4.2.2 * Improvement: include /usr/src/linux/.config file used to build the kernel. * Improvement: include ULi SATA support * Bug Fix: fixed problem with "swap-disable" configuration not being recognized.
July 10, 200817 yr Author Thanks for the help. The version (in my original post) is 4.2.4 pro. I'm going to keep a log here as I do each step: 1. Since the array is not started, I decided to click the "reboot" button to make sure we clear any "funky-ness" that I may have caused. 2. Wait for reboot... 3. Connect - main page comes up, says "try parity-swap"... 4. Click on "devices" - goes to "devices page" 5. Changed "disk2" to "unassigned" 6. Changed "parity" to "new disk" (300 GB PATA) 7. Changed "disk2" to "old parity disk" (250GB sata) 8. Clicked on "main" - goes to main page. 9. I see a "upgrade parity disk" option!!!! 10. I check "I'm sure i want to do this". 11. Clicked start. It worked! I can't explain why when I did this last time, it did not present me with the option. Perhaps I did it in a different order, or perhaps i changed it back and forth more than once. But, I even tried it a second time after rebooting... I must have done the same "bad thing" twice, because clearly your solution worked fine. I must have had "panic on" worrying about my data. Thank you very much. The 1TB drive is the spare for my other arrays. This array is the junior array with all the "hand me down" disks. I can't believe the forum fixed my problem in less than 20 minutes. Thank you all.
July 10, 200817 yr Author Just one more thought... Before the second reboot, I did do a: dd if=/dev/zero of=/dev/hda bs=2048k count=50 The 300GB pata drive had been part of an unraid array at one point and it did have a valid label and file system (reiserfs). The 'dd' was done in order to really convince unraid that this was a new drive and it was okay to write on it... I had run out of other ideas... I don't know if for sure if it was my "panic induced pilot error", or if this one thing did the trick. I tend to think "pilot error", but I wanted to mention it in case it helps someone else in the future... Thanks for the info about the new version of unraid. I should check for updates more often. The fact that I did not check is an indirect complement to the robustness of unraid - it worked so well for so long, that I forgot look for updates.
July 10, 200817 yr Just one more thought... Before the second reboot, I did do a: dd if=/dev/zero of=/dev/hda bs=2048k count=50 The 300GB pata drive had been part of an unraid array at one point and it did have a valid label and file system (reiserfs). The 'dd' was done in order to really convince unraid that this was a new drive and it was okay to write on it... I had run out of other ideas... I don't know if for sure if it was my "panic induced pilot error", or if this one thing did the trick. I tend to think "pilot error", but I wanted to mention it in case it helps someone else in the future... Thanks for the info about the new version of unraid. I should check for updates more often. The fact that I did not check is an indirect complement to the robustness of unraid - it worked so well for so long, that I forgot look for updates. I doubt if the "dd" to clear the 300Gig drive was the solution, but it did not hurt any. From your original post, you attempted to use the 300Gig drive in place of the failed disk2. That did not work, because if you were to do that, the parity drive would be smaller than a data drive. I can't tell what other combinations of assignments you tried, but I'm glad you are up and running now. There have been several improvements in the unRAID software, some only affect certain hardware, others, such as the "cache" drive feature apply only to the pro version. For you to upgrade to the latest release should only involve replacing the bzroot and bzimage files on the flash drive with those in the newest release. Make sure you know the disk assignments before you upgrade, that way, if Linux gets them out of order, you can assign them back to their original slots. Joe L.
Archived
This topic is now archived and is closed to further replies.