August 5, 201114 yr I have an unraid server 5.0 server, Beta 6a. I had installed 5 two TB disks as disk1, disk2, disk3, disk4 and disk5. No parity disk. Disk 5 had no data, so I wanted this to become a parity disk. I stopped the server, Cleared configuration and set the server up with parity disk and four data disks. But instead of setting disk 5 as parity i managed to set disk 1 as new parity. The parity sync started, and I found out what was wrong after about 10 seconds. I corrected the error, but now my disk1 shows as unformatted. Is there anyway I can salvage some of my data? I had all my photos on this disk. Edit: Found this link. http://lime-technology.com/forum/index.php?topic=1483.msg9973#msg9973 After what I can see, the disk is /dev/sdb. Do I need to change any of the parms for 5.0b6a?
August 6, 201114 yr Glad you found my old post. I don't think anything has changed about the reiserfsck commands you need to run. Just take your time. Good luck!
August 6, 201114 yr Just be careful to check that the disk is still /dev/sdb after reboots. The assignments can sometimes change, especially if other drives are added, removed, or moved between controller ports.
August 6, 201114 yr Author The job is now done. How do I access the data? I tried to set only disk1 active in UNRAID. Bu the disk is still "unformatted". did also try to mount the disk with: mount /dev/sdb /mnt/disk1 This gave me an error saying: mount: mount point /mnt/disk1 does not exist Here is the end of the joboutput: Flushing..finished Objects without names 4721 Empty lost dirs removed 40 Dirs linked to /lost+found: 520 Dirs without stat data found 32 Files linked to /lost+found 4201 Objects having used objectids: 490 files fixed 448 dirs fixed 42 Pass 4 - finished done 454674, 94 /sec Deleted unreachable items 7967 Flushing..finished Syncing..finished ########### reiserfsck finished at Sat Aug 6 19:20:52 2011 ########### root@Tower:/dev#
August 6, 201114 yr Did you reboot? I think I had to reboot before I saw anything if I remember right.
August 6, 201114 yr Author I have rebooted now, but the disk still shows as unformatted. Maybe I will try the job once more tonight? Thank you all for all the help so far.
August 6, 201114 yr Many users have run Reiserfsck for various problems including fixing accidentally reformatting and other corruptions similar to what you have done. I cannot remember ever a person reporting drive as unformatted at the end. Please double check you have performed all steps on the correct disk. Otherwise I am sorry but do not have another suggestion.
August 6, 201114 yr The job is now done. How do I access the data? I tried to set only disk1 active in UNRAID. Bu the disk is still "unformatted". did also try to mount the disk with: mount /dev/sdb /mnt/disk1 This gave me an error saying: mount: mount point /mnt/disk1 does not exist Here is the end of the joboutput: Flushing..finished Objects without names 4721 Empty lost dirs removed 40 Dirs linked to /lost+found: 520 Dirs without stat data found 32 Files linked to /lost+found 4201 Objects having used objectids: 490 files fixed 448 dirs fixed 42 Pass 4 - finished done 454674, 94 /sec Deleted unreachable items 7967 Flushing..finished Syncing..finished ########### reiserfsck finished at Sat Aug 6 19:20:52 2011 ########### root@Tower:/dev# The file system is on the first disk partition, /dev/sdb1 (note the "1" at the end of the device name.) The reiserfsck should have been run on /dev/sdb1, not on /dev/sdb. In the same way, to mount the drive you would mount it through the "md" device. (to keep parity in sync) Since you have no parity at this time, you can get away with mkdir /mnt/temp_drive mount -t reiserfs /dev/sdb1 /mnt/temp_drive Joe L. Joe L.
August 7, 201114 yr Author Thank you. SDB1 was correct. The disk is now readable. I found my pictures. about 50% was working, But now that I have the file names I can get many back from friends. The movies and series was 99% NOT working. But that is no problem. I can download them again. I hope a future release of Unraid will move paritydisk to the end of the list. As it is today it is far to easy for a newbe like me to make this mistake. Thank you all for your help.
August 7, 201114 yr Thank you. SDB1 was correct. The disk is now readable. I found my pictures. about 50% was working, But now that I have the file names I can get many back from friends. The movies and series was 99% NOT working. But that is no problem. I can download them again. I hope a future release of Unraid will move paritydisk to the end of the list. As it is today it is far to easy for a newbe like me to make this mistake. Thank you all for your help. Glad you at least had partial success, and have a path to regain all or most of your lost data. Your post made me remember the post below about reasons people lose data. Most (except #3) are still 100% accurate. Thought the latest round of users might benefit from reading this before getting bit. http://lime-technology.com/forum/index.php?topic=2601.msg21033#msg21033
August 7, 201114 yr Thank you. SDB1 was correct. The disk is now readable. I found my pictures. about 50% was working, But now that I have the file names I can get many back from friends. The movies and series was 99% NOT working. But that is no problem. I can download them again. I hope a future release of Unraid will move paritydisk to the end of the list. As it is today it is far to easy for a newbe like me to make this mistake. Thank you all for your help. Look in the lost&found directory, you may find a lot more of your files, but without their original names.
August 7, 201114 yr Author Yes, I have done that. All my files was in the lost+found directory. But most of them were corrupt. Many pictures did only have the first part of the original. Some had the first part correct and the second part was shifted to the right or to the left. And some only gave an error. I have copied all of them, and I will do the selection later on. I have shared many photos with my friends, so I think I will be able to get the most of them back. I also have some backups at work. It is just a lot of work getting all back to one collection. I have added a picture from my movie collection as an example. Most of my movies was also corrupt. Seems like data had been shifted around in the container. I tested many of them, but got errors in VLC. I think it will be easier to just download them once more than checking all. There may be errors later in the movie. I dont want to start looking on a movie that might start ok, and then fail in the middle. As long as I have all file names, it is easy to download again. I have a 50MB line for downloading. I know I should not be using a beta release for my pictures:(. I will now also always have a backup copy on another device of my important files. (I hope) Thanks again for all help.
August 8, 201114 yr I know I should not be using a beta release for my pictures:(. I will now also always have a backup copy on another device of my important files. (I hope) I'm glad to see you recovered some of your data. Your loss has nothing to do with the beta release however. I'm afraid that assigning a data disk as parity would clobber it in ANY raid system and any OS, regardless of version, beta or not. Joe L.
August 8, 201114 yr I know I should not be using a beta release for my pictures:(. I will now also always have a backup copy on another device of my important files. (I hope) I'm glad to see you recovered some of your data. Your loss has nothing to do with the beta release however. I'm afraid that assigning a data disk as parity would clobber it in ANY raid system and any OS, regardless of version, beta or not. Joe L. Although I do think running reiserfsck on the disk level device first (/dev/sdX), rather than the partition level (/dev/sdX1) device, reduced the effectiveness of the recovery. I updated my old "Data Recover Experience" post instructing to use the partition level device. The post actually instructs you to use the unRAID device (/dev/mdX) which works equally well and maintains parity. (It would be slower though if parity were in place.)
August 8, 201114 yr Author I know I should not be using a beta release for my pictures:(. I will now also always have a backup copy on another device of my important files. (I hope) I'm glad to see you recovered some of your data. Your loss has nothing to do with the beta release however. I'm afraid that assigning a data disk as parity would clobber it in ANY raid system and any OS, regardless of version, beta or not. Joe L. I did know that. I just wanted to point it out before anyone else told me what a fool I am keeping important data on beta software. This one was all MY fault. (My job is actually as an engineer working with SAN and disk systems. (So plase dont tell anybody of my blunder )). Just a few more questions. Can I format the drive to get rid of all the faulty data? I cannot find a way to format a formatted drive, other than first use it as a paritydisk for some seconds. Than it becomes unformatted. But this does not feel like the right way to do it. The disk has now 1.8TB of garbage. Is the disk as good as "newly formatted" (But with data) when I have completed the reiserfsck after my mistake? Should I just delete all the data and continue using the disk as it is? If I remove the disk completely can I than move disk2 to slot1, Disk3 to slot2 and so on after doing a reset config?. (Avoiding the Parity disk of course) Finaly, I can see that my disks dev name has different value: sdb, sdc, sde, and finaly hdc. All are connected to the internal sata contoller. It has a total of 6 sata connections. Is there any good reason for the "HDC" instead of sda to sdf. When i set the function of the SATA controller to IDE instead of raid, I can only see the first four drives. So I suspect it to be somting special with the last two connections and maybe that is the reason for the "hdc"?
August 9, 201114 yr Just a few more questions. Can I format the drive to get rid of all the faulty data? I cannot find a way to format a formatted drive, other than first use it as a paritydisk for some seconds. Than it becomes unformatted. But this does not feel like the right way to do it. The disk has now 1.8TB of garbage. Is the disk as good as "newly formatted" (But with data) when I have completed the reiserfsck after my mistake? Should I just delete all the data and continue using the disk as it is? You could just erase all the files... but if you wish to reformat, just issue (with the array started) umount /dev/mdX mkreiserfs /dev/mdX mount -t reiserfs /dev/mdX /mnt/diskX where mdX = md1 for disk1, md2 for disk2, etc. Joe L. (where If I remove the disk completely can I than move disk2 to slot1, Disk3 to slot2 and so on after doing a reset config?. (Avoiding the Parity disk of course) Finaly, I can see that my disks dev name has different value: sdb, sdc, sde, and finaly hdc. All are connected to the internal sata contoller. It has a total of 6 sata connections. Is there any good reason for the "HDC" instead of sda to sdf. When i set the function of the SATA controller to IDE instead of raid, I can only see the first four drives. So I suspect it to be somting special with the last two connections and maybe that is the reason for the "hdc"?
August 9, 201114 yr Finaly, I can see that my disks dev name has different value: sdb, sdc, sde, and finaly hdc. All are connected to the internal sata contoller. It has a total of 6 sata connections. Is there any good reason for the "HDC" instead of sda to sdf. When i set the function of the SATA controller to IDE instead of raid, I can only see the first four drives. So I suspect it to be somting special with the last two connections and maybe that is the reason for the "hdc"? sdX device names are used for SATA drives, whereas hdX device names are for PATA / IDE drives. Some BIOSes provide the option for SATA drives to emulate PATA drives. With this setting you wind up with hdX device names for SATA drives. My guess is that your 6 on-board SATA ports are actually provided by 2 different controllers - perhaps 4 via normal motherboard chipset, and 2 via onboard Jmicron controller. If so, looks like the Jmicron settings are set to emulate IDE. Might want to go through your BIOS settings closely and see.
August 9, 201114 yr Typically, drives that have been set to emulate IDE will be accessed more slowly than if run as SATA drives natively. This depends on the emulation in the BIOS, but many have found the performance as emulated IDE drives to be far lower than if configured as native SATA. The emulation exists because older windows OS (Win-XP and prior) have no native SATA drivers. Without the emulation mode, they would not be able to boot from the current SATA drives.
August 10, 201114 yr Author Okay. Then I understand a little bit more. My linux and Hw understanding is very limited. Thank you all for your help.
Archived
This topic is now archived and is closed to further replies.