October 27, 201312 yr Hello all, I just got into same situation like jetkins described in Replace failed drive with one larger than the parity drive? thread. My situation is following: - i'm currently runing unRaid 4.7 with 2TB drives - one 2TB drives failed (horrible noises during write -> disabled. It's at least readable) - i have 4x4TB drives on the way (ordered before failure) I planned to update to Unraid 5.x (because of >2TB support) and switch my parity with new 4TB drive, but this failure quite complicated things. On wiki and thread above it's described how to replace failed drive with one larger than parity (swap-disable) but there is only small note that upgrade to 5.x MAY be possible, nothing more. So my questions is Is it safe to upgrade unRaid 4.7 -> 5.x with one failed drive just before "swap-disable"? I don't have spare 2TB drive and don't want to buy new one to be able to make recovery before upgrade. With these 4 new drives I'll be on full capacity (controller, case) and so I don't want to buy new 2TB to stay in drawer when failed one comes back from reclamation. I have one additional question about "swap-disable" process. On wiki is written "Replace the parity hard disk with a new bigger one." and "Replace the failed hard disk with you old parity disk.". Does that mean that i must switch cables between failed disk and old parity (not sure if it's physically possible without rearranging :-S) to maintain "position" and by that unRaid recognize "swap-disable"? Or does unRaid recognize that new drive is larger than parity and do "swap-disable" automatically even without cable switching?
October 27, 201312 yr Hello all, I just got into same situation like jetkins described in Replace failed drive with one larger than the parity drive? thread. My situation is following: - i'm currently runing unRaid 4.7 with 2TB drives - one 2TB drives failed (horrible noises during write -> disabled. It's at least readable) - i have 4x4TB drives on the way (ordered before failure) I planned to update to Unraid 5.x (because of >2TB support) and switch my parity with new 4TB drive, but this failure quite complicated things. On wiki and thread above it's described how to replace failed drive with one larger than parity (swap-disable) but there is only small note that upgrade to 5.x MAY be possible, nothing more. So my questions is Is it safe to upgrade unRaid 4.7 -> 5.x with one failed drive just before "swap-disable"? I don't have spare 2TB drive and don't want to buy new one to be able to make recovery before upgrade. With these 4 new drives I'll be on full capacity (controller, case) and so I don't want to buy new 2TB to stay in drawer when failed one comes back from reclamation. I have one additional question about "swap-disable" process. On wiki is written "Replace the parity hard disk with a new bigger one." and "Replace the failed hard disk with you old parity disk.". Does that mean that i must switch cables between failed disk and old parity (not sure if it's physically possible without rearranging :-S) to maintain "position" and by that unRaid recognize "swap-disable"? Or does unRaid recognize that new drive is larger than parity and do "swap-disable" automatically even without cable switching? Personally, I think you should get the smaller size disk, then possibly sell it later if you want to upgrade. The odds of something going wrong is FAR too high based on your level of experience otherwise. Send lime-tech an e-mail for THEIR advice if you insist on upgrading to 5.0while in a degraded state and introducing a swap-disabled process in the mix. The failed drive is readable only because of parity protection. Be very cautious about doing anything un-needed. Best would be to copy the critical data off the failed drive now, to some other disk, even if it is an external USB drive. Then, set a new disk configuration without the failed disk, then upgrade to 5.0, then stop array, swap in new parity disk, start array, rebuild with the new 4TB drive as parity, then add the older 2TB ad a data drive, and lastly copy your data back to it from the external USB drive. Ask lime-tech to update the wiki for 5..0 on the swap-disabled process.
October 27, 201312 yr I would NOT complicate things by mixing the upgrade with the drive rebuild. If you want to avoid buying a 2TB drive (the best choice is to do that and replace the failed drive before doing the upgrade) ... you can do it as follows: (1) If you have good backups (as you should), simply make a directory of the failed disk (Use Directory Printer or any other nice directory tool you like); then just upgrade to v5, but do a "New Config" and assign the GOOD disks to the new configuration along with the new 4TB drives as parity and additional data drives; then Start the array and wait for the parity sync; and then copy the files that were on the failed disk to the array from your backups. (2) If you don't have current backups, then install one of the 4TB drives on a PC, initialize it and create a partition; and then copy the entire failed disk from your array to this disk. UnRAID will recreate the data for the disk from the other disks in the array (that's what fault tolerance is all about). THEN upgrade to v5 (as in #1), doing a New Config and putting the REST of the 4TB drives in the array (NOT the one you copied the data to). Then copy all of the data from the drive on your PC to the array => and THEN you can remove that drive from the PC; install it in the UnRAID box; pre-clear it; and add it to the array.
October 27, 201312 yr The 4.7 version of unRAID tracks "logical" disk assignments by the disk controller ports the physical disks were attached to. For swap-disabled to work, you must either have the disk controller port on the parity drive not change, or to do it on the disk assignment screen when the array is stopped. The 5.0 version tracks the disks by their serial numbers. It does not care about which port on a disk controller or cable is used. (You can't just swap cables, as it will still use them as previous when it finds them by their serial numbers) It is for that reason the swap-disabled procedure is needing update and clarification by Tom at lime-tech when on 5.0.
October 27, 201312 yr Yes, the details on swap-disable need to be clarified -- there have been quite a few questions from folks who have found that it did not work as they had hoped. Yet another reason to NOT do that in conjunction an upgrade. As both Joe and I noted, your best "solution" is to buy another 2TB drive and let the system rebuild the failed drive BEFORE you do the upgrade to v5. But the process I outlined will let you avoid doing that if you really don't want to buy one.
October 27, 201312 yr Author Thanks both for quick reaction. I quite expected such response, because i too consider it complicated and risky. So no surprise . First I'll try to look for some "spare" 2TB drive around me, but my chances there are slim. If not successful I'll do following (hope I understand you right): - copy data from failed disk (using other disks + parity) to one of new disks - shrink array by that one failed disk (not explicitly mentioned but i guess you mean shrinking of array) - check parity (to be sure) - update to 5.0 - switch parity to new 4TB disk - add all other new 4TB disks - copy backed data back to unRaid - check everything - add last 4TB which was used for backup @Joe L.: Thanks for explanation, I'll ask Tom by email about clarification (on wiki) how swap-disable works on 5.0.
October 27, 201312 yr As both Joe and I noted, your best "solution" is to buy another 2TB drive and let the system rebuild the failed drive BEFORE you do the upgrade to v5. But the process I outlined will let you avoid doing that if you really don't want to buy one. Or, short stroke a new 4TB to 2TB with an artificial HPA to accomplish the data rebuild, then do the upgrade, and expand the drive after you rebuild parity in 5.0 with the full size 4TB.
October 27, 201312 yr As both Joe and I noted, your best "solution" is to buy another 2TB drive and let the system rebuild the failed drive BEFORE you do the upgrade to v5. But the process I outlined will let you avoid doing that if you really don't want to buy one. Or, short stroke a new 4TB to 2TB with an artificial HPA to accomplish the data rebuild, then do the upgrade, and expand the drive after you rebuild parity in 5.0 with the full size 4TB. I would do that; you would do that; but I don't think this is something I'd suggest a relative novice attempt
October 27, 201312 yr ... I'll do following (hope I understand you right): - copy data from failed disk (using other disks + parity) to one of new disks - shrink array by that one failed disk (not explicitly mentioned but i guess you mean shrinking of array) - check parity (to be sure) - update to 5.0 - switch parity to new 4TB disk - add all other new 4TB disks - copy backed data back to unRaid - check everything - add last 4TB which was used for backup NO. Not exactly right. To be a bit more explicit, do the following: - copy data from failed disk (using other disks + parity) to one of new disks -- Look at the Web GUI and WRITE DOWN the serial numbers of all of your data drives (note which one is the bad drive) After that you do NOT need to "shrink the array", "check parity", or anything else with the v4.7 system. Just update the flash drive to v5.0; install your 4TB drive for parity; boot to v5; go to Utils and select "New Config"; assign all of your good data drives (NOT the failed drive), and assign the 4TB drive as parity; then Start the array and let it do a parity sync. Be sure your drive assignments are correct (that's why I had you write down all the serial numbers). After it does a parity sync, do a parity check to confirm all went well. NOW you're ready to add the other 4TB drives; copy the data back from the one you saved it to; and finally add that last 4TB drive to the array.
Archived
This topic is now archived and is closed to further replies.