Roger Posted August 20, 2011 Author Share Posted August 20, 2011 Drive 8 comes up unformatted, should I bang it again with a format? Link to comment
Roger Posted August 20, 2011 Author Share Posted August 20, 2011 Joe L, since we did it before, figured, why not hit Disk 8 with a format again and run this command. We tried it before, and it's running, crashed early on, I'll let you know. reiserfsck --rebuild-tree -S --badblocks /boot/badblocks_disk8 /dev/sdX1 Don't suppose you found any more information on the bad block solution. Disk 8 is the disk that was being formatted when I shut it down. Don't have another 1.5 tera drive, have a 1 handy. I know we are not there yet, just letting you know. I know this is a stupid question, but; I found this command that said it fixes bad sectors, fsck -y /dev/sdaX, could it work? Roger Link to comment
Joe L. Posted August 20, 2011 Share Posted August 20, 2011 Drive 8 comes up unformatted, should I bang it again with a format? Sure. Then, the SMART firmware should re-allocate blocks it knows are un-readable. before you do, try this command to get a SMART report: smartctl -a /dev/sdo | grep -i alloc I'm interested in the last number on eack line. Then, you can format the drive using the button on the web-interface, then, from the command line, run reiserfsck --rebuild-tree -S /dev/md8 to see what it can recover. Link to comment
Roger Posted August 20, 2011 Author Share Posted August 20, 2011 Before I got your message, started your 3rd badblock rebuild command that has the /boot/badblocks_disk8. I locked at the link and I see how, rather than use a floppy, it should be using the badblock file we created that's on boot, to rebuild. It's at 2348952185, usually fails at about 1948942185'sh. So, I'll know soon whether this works, if not, I'll try your other suggestion. Roger Link to comment
Roger Posted August 20, 2011 Author Share Posted August 20, 2011 Have the array stopped, seems to work better that way. I did make the mistake a couple of times using the /dev/sdx command before I stopped the array, then I hit Control C, then I stopped, and ran again. Did I hurt anything? Link to comment
Joe L. Posted August 21, 2011 Share Posted August 21, 2011 Have the array stopped, seems to work better that way. I did make the mistake a couple of times using the /dev/sdx command before I stopped the array, then I hit Control C, then I stopped, and ran again. Did I hurt anything? You should be fine. Can't do much more damage at this point... (with no parity drive assigned, can't mess up parity, as it is messed up regardless until you assign it and re-calculate it. Link to comment
Roger Posted August 21, 2011 Author Share Posted August 21, 2011 Joe L, O.K., last number on first line is 926, last number on second line is 194. So, what's next? Roger Link to comment
Roger Posted August 21, 2011 Author Share Posted August 21, 2011 Disk 8 failed again on rebuild at 192133798 Link to comment
Roger Posted August 21, 2011 Author Share Posted August 21, 2011 Stuck, this look interesting. Canget it to work, but; maybe I'm not entering the command correctly. http://smartmontools-support-list.2498698.n2.nabble.com/ReiserFS-Badblocks-Script-td5307380.html Link to comment
Roger Posted August 21, 2011 Author Share Posted August 21, 2011 :'( Joe L, hope you didn't give up on me. Waiting for your next solution Link to comment
Joe L. Posted August 21, 2011 Share Posted August 21, 2011 :'( Joe L, hope you didn't give up on me. Waiting for your next solution Do you have a free slot in your array to plug in the spare drive? (more accurately, a free physical slot/disk controller port) If yes, we can get the rest of your array up and running, get parity established on it, leave disk8 out of the array for now and come back to it. I'll be away from the computer most of tomorrow, so just run the --rebuild-tree -S on all the other drives you have not, then un-assign disk8, assign parity, and get the bulk of your disks protected once more. If you have the spare drive, put in into the server, do not assign it as of yet. Joe L. Link to comment
Roger Posted August 21, 2011 Author Share Posted August 21, 2011 Unfortunately, no spare slots, I'm fully, but; I'll unassign 8, and set parity in the morning. Just two disks to go and I've run the rebuild on all. THanks for your help. In my gut, I know there must be a way to fix Disk8, just not understanding how. ROger Link to comment
Roger Posted August 22, 2011 Author Share Posted August 22, 2011 Joe L, except for Disk8, all seems to be now working. One terabyte drive, so I know there is stuff on there, I would want. Electricity just went out on my house when the parity drive was 70% completed. Some issue with the neighborhood. Does this cause the 15 hours of work to get to 70% complete to go poof on the parity drive. WIll this potentially cause all the work we did so far to fail? Roger Link to comment
Roger Posted August 23, 2011 Author Share Posted August 23, 2011 Joe L, things seem to be working, 60% done building parity on the parity drive, after the power went out and it had to restart. Look forward to trying to save Disk8 Roger Link to comment
Joe L. Posted August 23, 2011 Share Posted August 23, 2011 It figures, you would lose power in the middle of building parity. I"m several hundred miles from my own servers. (on a trip to the beach) Hoping hurricane Irene does not hit before I leave for home later this week. We will need it learn how it force write the bad block on disk8 it allow reiserfsck it complete. Posting from Android based e-reader. Joe L. Link to comment
Roger Posted August 24, 2011 Author Share Posted August 24, 2011 Joe L, know your busy, just checking to see if you have any advice on Disk8, and how to recover the data on it. Tried to run a regular rebuild on Disk8 jsut to see what it would do. Same location it fails 192133789, same message, Cannot read the block (52050987) Roger Link to comment
Joe L. Posted August 24, 2011 Share Posted August 24, 2011 You had mentioned you had nearly 1000 sectors that had already been reallocated And almost 200 pending reallocation when they are next written. We need to learn how to congert the faling block numbr reported by reiserfs int o one we can manually write it force a reallocation. Link to comment
Roger Posted August 24, 2011 Author Share Posted August 24, 2011 Joe L, the link you provided made sense to do this. Once the badblock file was created. And we did that with a list. Then the reiseefschk command in step 2 of 3 would reallocate or write over badblocks. Then 3rd rebuild tree. Made sense but for whatever reason Linux doesn't recognize the command. I saw some other references to how this might be done manually one block at a time but really can't understand it well enough to translate to our system. Roger Link to comment
Roger Posted August 24, 2011 Author Share Posted August 24, 2011 Joe L, don't know what congert means, but; will wait to have you explain it to me. Roger Link to comment
Joe L. Posted August 24, 2011 Share Posted August 24, 2011 Sorry.. Congert=correct when the auto correct on my phone goes bananas Link to comment
prostuff1 Posted August 24, 2011 Share Posted August 24, 2011 Sorry.. Congert=correct when the auto correct on my phone goes bananas not as bad as some of these. Link to comment
Roger Posted August 25, 2011 Author Share Posted August 25, 2011 Joe L; any more progress on a way to fix Disk 8? Link to comment
Joe L. Posted August 25, 2011 Share Posted August 25, 2011 Joe L; any more progress on a way to fix Disk 8? I'm sevreal hundred miles from my PC and server. Real tough to research on this phone I am using to respond. Do you have parity established again? Link to comment
Roger Posted August 25, 2011 Author Share Posted August 25, 2011 Yes, parity is established Link to comment
Joe L. Posted August 25, 2011 Share Posted August 25, 2011 Yes, parity is established good. Now we will focus on disk8 via the sdX1 device.(unless it is assigned to the array, if it is we use md8 to get to it. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.