thany Posted November 26, 2022 Share Posted November 26, 2022 (edited) So I did the following: 1. Pull the old drive that is to be replaced. 2. Insert the new one 3. Stop the array 4. Select the new drive as part of the array 5. Start the array, letting it rebuild data No shutdowns in between, as this computer has hotswap bays, and a controller that supports it. So all goes well. All data is still there. Just how it's supposed to work. Excellent Only the shares to that data are now gone Is this intentional? How do I get them back? Or rather, how do I not lose them the next time I replace a drive? Is it safe to just recreate the shares, or will this have strange consequences? I'm on 6.11.5, and my array consists of 4 drives, using 1 for parity. Edited November 26, 2022 by thany Quote Link to comment
JorgeB Posted November 26, 2022 Share Posted November 26, 2022 Please post the diagnostics. Quote Link to comment
trurl Posted November 26, 2022 Share Posted November 26, 2022 No point in hotswapping array or pool devices since nothing will be done with a new disk until you stop the array and assign it. Regardless, shares should be there if the data is there, simple as that. What evidence do you have that the data is there? Attach diagnostics to your NEXT post in this thread. Quote Link to comment
trurl Posted November 26, 2022 Share Posted November 26, 2022 3 hours ago, trurl said: shares should be there if the data is there, simple as that To elaborate, user shares are simply the combined top level folders on array and pool. If you create a user share in the webUI, a top level folder named for that share is created on array or pool as needed according to the settings for the share. If you create a top level folder on array or pool, it is automatically part of a user share named for that folder. Any user share you have not made settings for will have default settings. If those top level folders exist, user shares named for those folders should exist. Possibly you have some other problem such as unmountable disks that is keeping user shares from working. 4 hours ago, thany said: Is it safe to just recreate the shares Should be unnecessary and unlikely to help. Diagnostics would probably tell us what the problem is and how to proceed. Quote Link to comment
thany Posted November 26, 2022 Author Share Posted November 26, 2022 Sure, here it is. It's not because it's still rebuilding, right? Because I feel data should be accessible even during a rebuild. unraid-diagnostics-20221126-2302.zip Quote Link to comment
trurl Posted November 26, 2022 Share Posted November 26, 2022 23 minutes ago, thany said: data should be accessible even during a rebuild Assuming everything is working correctly it is accessible. Filesystem Size Used Avail Use% Mounted on /dev/md3 15T 127G 15T 1% /mnt/disk3 Emulated disk3 is mounted but seems mostly empty. Did you perhaps format it? Lots of stuff in syslog about zfs and luks (encryption) but none of your array or pools seem to be using either. However, some of the zfs entries seems related to sdk which is currently disk3 and xfs. I don't use zfs or luks so not sure what that is all about. What more can you tell us about all that? Quote Link to comment
thany Posted November 26, 2022 Author Share Posted November 26, 2022 (edited) 15 minutes ago, trurl said: Emulated disk3 is mounted but seems mostly empty. Did you perhaps format it? There's just not a whole lot of data on the array yet. The thing is, I accidentally bought SMR disks and am now replacing all of them. This seemed like a good enough oppertunity to, instead of starting over with nothing, do the disk replacement procedures just to see if it all works the way I think it will work when a disk *actually* fails. 15 minutes ago, trurl said: Lots of stuff in syslog about zfs and luks (encryption) but none of your array or pools seem to be using either. However, some of the zfs entries seems related to sdk which is currently disk3 and xfs. I don't use zfs or luks so not sure what that is all about. What more can you tell us about all that? I'm thinking the zfs stuff is probably unrelated. I've got a secondary zfs array on 4 separate disks, setup using that zfs plugin and zfs master. I have set up a share to that secondary array, and when stopping the main array, unRAID desperately wants to "stop" that share as well, which doesn't work because it's not on the array. I'm guessing that's what you're noticing. I did this by simply creating a symbolic directory link to the data in /mnt/user, which has also seemingly been deleted. Edited November 26, 2022 by thany Quote Link to comment
trurl Posted November 26, 2022 Share Posted November 26, 2022 I'm guessing your customization has broken user shares, and in fact, /mnt/user is missing as you say. If you reboot in SAFE mode (which will of course not use the ZFS plugin) user shares will probably work again, then maybe you can get to the bottom of these other problems. Might see the Support thread for that plugins to see if there is anything there for you. Quote Link to comment
thany Posted November 27, 2022 Author Share Posted November 27, 2022 (edited) I really doubt it's the ZFS plugin messing it up. I just rebooted the server (normally) and the shares are back. Even the one I added manually. Could this be a bug in unRAID? Edited November 27, 2022 by thany Quote Link to comment
trurl Posted November 27, 2022 Share Posted November 27, 2022 In order to report a bug you have to confirm it still happens in SAFE mode. If it doesn't happen in SAFE mode you need to figure out which plugin is causing the problem and report it to that plugin's thread. Quote Link to comment
thany Posted November 27, 2022 Author Share Posted November 27, 2022 I can't know if it happens in safe mode, because for that I have to reboot into it. But rebooting normally also makes the shares reappear. Ergo, safe mode is not going to reveal anything we don't already know. Just tried with the next disk to replace. Quote Link to comment
thany Posted November 28, 2022 Author Share Posted November 28, 2022 (edited) Sorry what? I've replaced the last disk to be replaced, and now after that neccesary reboot the network settings are gone? eth0 is back on DHCP, by which it manages to NOT even get an IP-address but one of those useless 169.* autogenerated addresses. Why has it done that?? Took me another half hour of running back and forth, figuring out how to edit the network.cfg (a good reference is missing, can someone please make sure documentation is complete?) And discovered that GUI mode is completely not-working, even though it can absolutely output a video signal, it doesn't get any further than a blinking cursor in an otherwise blank screen. Is there any documentation on that, while we're at it? Edited November 28, 2022 by thany Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.