June 10, 200818 yr Yesterday i added 2 new drives to my array and everything went fine. Last night i powered down the server and went to sleep. When i powered up the server this morning i noticed the server had started a parity sync. The webapp was reporting a invalid parity. This had me worried for a sec, so i decided to check my data to make sure everything was ok. Well it wasn't. Three of my folders were empty. I fired up telnet and checked to see if maybe something went wrong with the array and sure enough my data was still on the drive. The data is there, but I'm unable to access any of it due to a permissions problem. root@Tower:/mnt/disk2/music/hip-hop/Hip-Hop# cd VA-The_Bassment -bash: cd: VA-The_Bassment: Permission denied basically my data is there, yet I'm unable to see it on the disk share and through the user share. The only was i can see the data is through telnet
June 10, 200818 yr That's unRaid protecting you from listening to Hip-Hop Sorry, probably not funny to you. How did you copy the files to the share to begin with?
June 10, 200818 yr Author The disk that is having the problem is not one of the new disks added. The array was working perfectly prior to adding the new drives. Everything was copied over from a windows box. i restarted and captured the syslog...
June 10, 200818 yr You may need to unmount the drive and do a reiserfsck on the drive. If you login as root, you should be able to access anything. One of the directory pointers may have an issue. See if there is data in a lost+found directory. Wait before doing anything further until people have a look at your syslog.
June 11, 200818 yr I agree with WeeboTech, you need to run reiserfsck on Disk 2 (md2, sdd, Seagate 400MB). See the instructions here (use md2): Check Disk Filesystems. Although the syslog you captured is helpful (it shows the problem with Disk 2), it would have been better to capture a syslog *before* you restarted. When you restart, all of the current error messages are lost, since the entire operating system and the syslog are held only in memory. Without those error messages, we can't even speculate on what went wrong. It is possible there was a crash or power off in the middle of a write to Disk 2, or memory corruption that affected the reiserfs control structures associated with Disk 2.
June 11, 200818 yr Author Here are the results of the test.. Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs. Bad nodes were found, Semantic pass skipped 7 found corruptions can be fixed only when running with --rebuild-tree ########### reiserfsck finished at Tue Jun 10 13:13:57 2008 ########### root@Tower:~# what happens if i rebuild tree since my parity seems to be bad also? will the data stay intact during this process?
June 11, 200818 yr Don't worry about the parity drive until you have recovered your data, by running reiserfsck, with whatever parameters it suggests. Once the data drive(s) are fine, then run a parity check, and it will correct any incorrect parity bits. Continue with the reiserfsck instructions on that page, with the following command: reiserfsck --rebuild-tree /dev/md2 Afterward, as was suggested above, check for a lost+found folder on Disk 2, for fragments of files that could not be completely recovered. Although probably unnecessary, I like to run one last reiserfsck, just to verify that everything is clean. That would be reiserfsck /dev/md2. Then remount and restart samba and run your parity check.
June 11, 200818 yr I did the rebuild tree thing a while back. My disk had been completely hosed - all the record-keeping areas at the front of the drive were overwritten with binary 0s! It worked but took hours and hours. Don't expect it to be a 5 minute thing. I needed to bring down the array and bring it back up (I think I actually did a full reboot) in order to see the corrected contents. Good luck!
June 13, 200818 yr If your unRAID server is not on a UPS, you might want to put it on one. The only time my unRAID server starts a parity check on its own is if I had a power interruption and it was not on a UPS. Sorry to learn you did not recover all your files. Joe L.
Archived
This topic is now archived and is closed to further replies.