local.bin Posted September 13, 2012 Share Posted September 13, 2012 I was running 5rc2 and had cnib problems today, tried to sort via web and failed, so updated to latest rc6 beta, hoping it was a bug and had been resolved (maybe) Had I/O errors on drive 3 (hdb) which effectively failed. Although seemed fine before the update. Parity is showing green as well as other disks showing green. I rebooted ( ) and it said there was an unformatted drive. I formatted ... disk is red. I went to try and initiate rebuild from parity by deselecting and reselecting disk, but the disk show as 'Not present' even though the bios sees it and it previously formatted by unraid?? Can I still rebuild it from parity? If so what next please Log attached after reboot, sorry. syslog.txt.zip Link to comment
jumperalex Posted September 13, 2012 Share Posted September 13, 2012 Abuse: you did WAY too many things blindly Advice: When something works, then suddenly stops working, "updating" is never the solution until you are sure that today was the day your system ran across an edge case scenario that is addressed by the update. Useful Advice: You're goint to want to start checking connections and cables and then wait for someone else on here to chime in with the correct reiserfs scan command. Link to comment
local.bin Posted September 13, 2012 Author Share Posted September 13, 2012 Abuse accepted Not sure a reiser check is going to help me, as the disk has been formatted again. I dont want to touch anything without some guidance I was hoping I could get a trust parity recover going ... Link to comment
Joe L. Posted September 13, 2012 Share Posted September 13, 2012 Abuse accepted Not sure a reiser check is going to help me, as the disk has been formatted again. I dont want to touch anything without some guidance I was hoping I could get a trust parity recover going ... You are way past that as a possibility. When you re-formatted, it updated parity. The only possibilities are: 1. See if you can mount the physical disk's first partition using a third party tool. (un-assign it from the array first) It might not be accessible, but if the BIOS see it, you might be in luck. If it really died, and that was why it could not be written, that will not work. 2. Use the reiserfsck --rebuild-tree -S /dev/md3 option to reiserfsck to attempt to recover what it can from your re-formatting of the /dev/mdX device. Joe L. Link to comment
local.bin Posted September 13, 2012 Author Share Posted September 13, 2012 :'( Ok thanks. The disk appears fine, well it passed a short smartcrtl test. Shows a previous error waking up from sleep. Where do I run option 2, from unraid telnet? If I can't get the data, how do I get the disk back into the array? Thanks Link to comment
jumperalex Posted September 13, 2012 Share Posted September 13, 2012 Abuse accepted Not sure a reiser check is going to help me, as the disk has been formatted again. I dont want to touch anything without some guidance I was hoping I could get a trust parity recover going ... You are way past that as a possibility. When you re-formatted, it updated parity. The only possibilities are: 1. See if you can mount the physical disk's first partition using a third party tool. (un-assign it from the array first) It might not be accessible, but if the BIOS see it, you might be in luck. If it really died, and that was why it could not be written, that will not work. 2. Use the reiserfsck --rebuild-tree -S /dev/md3 option to reiserfsck to attempt to recover what it can from your re-formatting of the /dev/mdX device. Joe L. So it looks like the array is currently ONLINE .. is that correct? Will he want to restart the array in Maintenance Mode? Link to comment
local.bin Posted September 13, 2012 Author Share Posted September 13, 2012 Array is stopped at the moment, trying to source a machine to read this disk .... Link to comment
local.bin Posted September 13, 2012 Author Share Posted September 13, 2012 Is it possible the disk went offline again before the format was actioned? I went to put the disk 3 back in the array and go into maintenance mode, where it said the disk was replaced and a data rebuild was to be done. There are alot of writes going on for the disk 3, does that mean data is being written back to it, or does that happen regardless? Damn I'm such a noob at this!! Link to comment
Joe L. Posted September 13, 2012 Share Posted September 13, 2012 Is it possible the disk went offline again before the format was actioned? Yes, in which case the physical disk was not reformatted, but the "simulated" one was. (in other words, parity was updated to reflect the formatting) I went to put the disk 3 back in the array and go into maintenance mode, where it said the disk was replaced and a data rebuild was to be done. There are alot of writes going on for the disk 3, does that mean data is being written back to it, or does that happen regardless? Yes, but from your description, what is being written is the "formatted" drive. Damn I'm such a noob at this!! yes, as we all are. but every situation is different... and all it take is one "aw-shoot" moment to lose data. unRAID's parity protection is not a backup of your data. It is protection against a physical disk failure, but only if you act appropriately when a disk fails. When dealing with a disk failure "formatting" is almost never the correct action. Link to comment
local.bin Posted September 13, 2012 Author Share Posted September 13, 2012 Yes data seems to be writing back to the previously 'formatted' drive (now seen as a replaced drive). I just don't see how that would be if parity didn't have that data to recover; or have I misunderstood how the data recreation works. I thought formatting would allow the disk to be recreated because I don't have another 2tb drive to replace it. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.