deva1983 Posted March 26, 2013 Share Posted March 26, 2013 Hi, I added an extra SATA controller to my ESXI box. For some reason, this changed the existing 'address' of the already assigned controller to unRAID. Result: unraid said the config was invalid. So I did an initconfig, and reassigned the drives. Now it says the 2 data drives are unformatted ?! Is there any way I can restore some data ? Can somebody plz help me ? Quote Link to comment
Joe L. Posted March 26, 2013 Share Posted March 26, 2013 Hi, I added an extra SATA controller to my ESXI box. For some reason, this changed the existing 'address' of the already assigned controller to unRAID. Result: unraid said the config was invalid. So I did an initconfig, and reassigned the drives. Now it says the 2 data drives are unformatted ?! Is there any way I can restore some data ? Can somebody plz help me ? Not without more detail. You must attach a syslog for analysis in most cases. Odds are the disks could not be mounted. If, by some reason you messed with the partitioning of the disks,then they will need to be set back as they were. (I hope you remember if they were 4k-alligned, or not. If not, you'll have to verify first before going off trying to fix a file-system where is not currently located.) What version of unRAID? To unRAID, any disk it cannot mount the first partition as a reiserfs is "unformatted" Your first step is to check if the file system on the disk is OK. Instructions in the wiki. You'll be running reiserfsck on /dev/md1, /dev/md2, etc, based on the disk not mountable. Quote Link to comment
deva1983 Posted March 26, 2013 Author Share Posted March 26, 2013 thanks for your quick reply. When I follow the wiki, /dev/md1 is not found. The gui shows an invalid configuration (Too many wrong and/or missing disks!), it seems like it replaced disk1 and the parity disk. Disk2 is green. Should i run initconfig ? Quote Link to comment
Joe L. Posted March 26, 2013 Share Posted March 26, 2013 thanks for your quick reply. When I follow the wiki, /dev/md1 is not found. The gui shows an invalid configuration (Too many wrong and/or missing disks!), it seems like it replaced disk1 and the parity disk. Disk2 is green. Should i run initconfig ? If you run initconfig, you will immediately invalidate parity. (and probably erase any chance of re-constructing data from it and the other data disks) Without the syslog for analysis, nobody can give you much more advice. You've not even bothered to answer the most basic question I asked. What version of unRAID are you running? Which disks are "unformatted" (not able to be mounted) What do you mean unRAID "replaced" disk1 and the parity disk? Joe L. Quote Link to comment
deva1983 Posted March 26, 2013 Author Share Posted March 26, 2013 Sorry, I'm a bit stressed here I'm using 4.7. disks atm are not reported as unformatted anymore (see attached image), but it appears as if it has swapped parity and disk 1. parity should be EZRX syslog.txt Quote Link to comment
deva1983 Posted March 26, 2013 Author Share Posted March 26, 2013 Maybe thisis a solution ? do a initconfig, and don't assign a parity disk to check if I selected the right parity disk ? md2 reiserfs says it can't find the superblock Quote Link to comment
deva1983 Posted March 26, 2013 Author Share Posted March 26, 2013 BTW, if I unassign the parity disk, and assign the disks it 'expects' to disk1 and 2, I get 2 unformatted drives. EDIT: removed disks and inserted in another PC, but was unable to mount all 3 of them :'( syslog.txt Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 Do you know for sure which is your parity disk? (by serial number?) Since you put the drives on a different controller, it is a 50/50 chance they would be on opposite ports and be swapped. I'd put them back as they should have been, even though they seem "unformatted" (unable to be mounted) Do you remember if you formatted them originally as 4k-alligned (starting on sector 64), or un-alligned (starting on sector 63) Joe L. Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 Run this on each of the two data drives and post the output dd if=/dev/sdX count=195 | od -c -A d | sed 30q (where sdX = sda, sdb, etc as appropriate for your physical disks) also need to see the output of fdisk -lu /dev/sdX for each of the drives Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Do you know for sure which is your parity disk? (by serial number?) Since you put the drives on a different controller, it is a 50/50 chance they would be on opposite ports and be swapped. I'd put them back as they should have been, even though they seem "unformatted" (unable to be mounted) Do you remember if you formatted them originally as 4k-alligned (starting on sector 64), or un-alligned (starting on sector 63) Joe L. I'll try this when I get back from home. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Run this on each of the two data drives and post the output dd if=/dev/sdX count=195 | od -c -A d | sed 30q (where sdX = sda, sdb, etc as appropriate for your physical disks) also need to see the output of fdisk -lu /dev/sdX for each of the drives I believe disk sdd and sda to by my data disks (sda being empty). dd sda: 0000000 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000448 \0 \0 203 \0 \0 \0 @ \0 \0 \0 / e p t \0 \0 0000464 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000496 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 U 252 0000512 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0099840 dd sdd 0000000 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000432 \0 \0 \0 \0 \0 \0 \0 \0 240 333 371 353 \0 \0 \0 \0 0000448 \0 \0 203 \0 \0 \0 @ \0 \0 \0 p 210 340 350 \0 \0 0000464 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000496 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 U 252 0000512 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0098304 240 034 222 023 350 325 \ \0 027 266 004 \0 \0 \0 \0 0098320 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 030 336 312 A 0098336 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098352 006 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098368 \0 \0 \0 \0 \a \0 $ ' \0 \0 \0 \0 342 205 \0 \0 0098384 \0 \0 \0 \0 301 202 336 $ 252 F \n 260 6 ; @ [ 0098400 026 0 8 347 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098416 \0 \0 \0 \0 \t \0 \0 \0 342 021 z \0 \0 \0 \0 \0 0098432 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0098512 \b 245 \0 \0 020 245 \0 \0 022 245 \0 \0 020 245 \0 \0 0098528 022 245 \0 \0 210 240 \0 \0 262 240 \0 \0 210 240 \0 \0 0098544 262 240 \0 \0 210 240 \0 \0 262 240 \0 \0 210 240 \0 \0 * 0099840 sdc ('presumed' parity) 0000000 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000432 \0 \0 \0 \0 \0 \0 \0 \0 244 333 371 353 \0 \0 \0 \0 0000448 \0 \0 203 \0 \0 \0 @ \0 \0 \0 p 210 340 350 \0 \0 0000464 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0000496 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 U 252 0000512 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0098304 240 034 222 023 350 325 \ \0 027 266 004 \0 \0 \0 \0 0098320 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 030 336 312 A 0098336 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098352 006 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098368 \0 \0 \0 \0 \a \0 $ ' \0 \0 \0 \0 342 205 \0 \0 0098384 \0 \0 \0 \0 301 202 336 $ 252 F \n 260 6 ; @ [ 0098400 026 0 8 347 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 0098416 \0 \0 \0 \0 \t \0 \0 \0 342 021 z \0 \0 \0 \0 \0 0098432 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 \0 * 0098512 \b 245 \0 \0 020 245 \0 \0 022 245 \0 \0 020 245 \0 \0 0098528 022 245 \0 \0 210 240 \0 \0 262 240 \0 \0 210 240 \0 \0 0098544 262 240 \0 \0 210 240 \0 \0 262 240 \0 \0 210 240 \0 \0 * 0099840 fdisk sda: Disk /dev/sda: 1000.2 GB, 1000203804160 bytes 1 heads, 63 sectors/track, 31008302 cylinders, total 1953523055 sectors Units = sectors of 1 * 512 = 512 bytes Disk identifier: 0x00000000 Device Boot Start End Blocks Id System /dev/sda1 64 1953523054 976761495+ 83 Linux Partition 1 does not end on cylinder boundary. fdisk sdd Disk /dev/sdd: 2000.3 GB, 2000398934016 bytes 1 heads, 63 sectors/track, 62016336 cylinders, total 3907029168 sectors Units = sectors of 1 * 512 = 512 bytes Disk identifier: 0xebf9dba0 Device Boot Start End Blocks Id System /dev/sdd1 64 3907029167 1953514552 83 Linux Partition 1 does not end on cylinder boundary. fdisk sdc ('presumed' parity) Disk /dev/sdc: 2000.3 GB, 2000398934016 bytes 1 heads, 63 sectors/track, 62016336 cylinders, total 3907029168 sectors Units = sectors of 1 * 512 = 512 bytes Disk identifier: 0xebf9dba4 Device Boot Start End Blocks Id System /dev/sdc1 64 3907029167 1953514552 83 Linux Partition 1 does not end on cylinder boundary. Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 It appears as if there is no evidence of the "reiserfs superblock" on any of your disks. Apparently, it has been overwritten. I expected output like this: http://lime-technology.com/forum/index.php?topic=23246.msg205459;topicseen#msg205459 containing a "R e I s E r 2 F s " string in the output. Let's assume you are correct in identification of the "empty" disk as "sda" If you were just now adding the third drive as a second data disk, then the parity disk and the first data disk would actually be mirrors of each other since even parity is used in unRAID. Assuming you agree that you specified 4k-alligned when setting up your array, and your partitions all start on sector 64 (it is what fdisk -lu reports) then you need to rebuild the superblock on the affiliated /dev/mdX device (disk1 = /dev/md1, disk2 = /dev/md2 ). reiserfsck --rebuild-sb /dev/mdX It needs specific responses to its prompts, and they are not the defaults, so beware.... See here for the correct responses: http://lime-technology.com/forum/index.php?topic=25132.msg219925;topicseen#msg219925 Joe L. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Thanks for all your help you ! As I searched the forums, I also noticed that i missed the REISER part in the dd. I'm pretty sure (99%) which is the parity disk (as I took the fastest drive), but what if I am mistaken ? A rebuild-sb would destroy the parity I suppose ? Is there any way to make sure which is the parity disk ? I'll try your instructions tonight. Thanks again ! Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 Thanks for all your help you ! As I searched the forums, I also noticed that i missed the REISER part in the dd. I'm pretty sure (99%) which is the parity disk (as I took the fastest drive), but what if I am mistaken ? A rebuild-sb would destroy the parity I suppose ? Is there any way to make sure which is the parity disk ? I'll try your instructions tonight. Thanks again ! Did you only have two disks (one parity, one data) and were in the process of adding a third disk? (the one you say is completely empty and is expected to be unformatted, since you've never formatted it?) Joe L. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Well, the complete story: I started with 2 disks, then early last week I purchased an extra SATA controller and disk (1GB). Installed both, cleared the disk and added it to the array. Everything OK. However, yesterday I was experiencing slow read speeds (8 MB/s compaired to 30 the week before), so I suspected the new controller gave conflicts. I removed the controller (and disk), which (in ESXI) also changed the address of my other (onboard) controller ... Re-passed-through the onboard controller, re-assigned the disks in unraid and started the array with 1 disk missing (the one of 1TB). However, the slow speeds remained, so I reseated the PCIx SATA controller and disk. The address of the onboard controller changed again, so I had to re-assign them both again to unRAID. Then the shit hit the fan, and it detected unformatted disks (which might have occurred after a parity sync (which I aborted quickly)?)... Now it looks like it confused the previous parity disk with disk1... Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 Well, the complete story: I started with 2 disks, then early last week I purchased an extra SATA controller and disk (1GB). Installed both, cleared the disk and added it to the array. Everything OK. However, yesterday I was experiencing slow read speeds (8 MB/s compaired to 30 the week before), so I suspected the new controller gave conflicts. I removed the controller (and disk), which (in ESXI) also changed the address of my other (onboard) controller ... Re-passed-through the onboard controller, re-assigned the disks in unraid and started the array with 1 disk missing (the one of 1TB). However, the slow speeds remained, so I reseated the PCIx SATA controller and disk. The address of the onboard controller changed again, so I had to re-assign them both again to unRAID. Then the shit hit the fan, and it detected unformatted disks (which might have occurred after a parity sync (which I aborted quickly)?)... Now it looks like it confused the previous parity disk with disk1... In that case, since you really have no clue which disks was the parity disk, I'd try a --rebuild-sb on both and not assign any parity disk for now. (there is one disk with no data, so let's figure that one as the new disk you were attempting to add) We'll deal with it last since it has nothing on it. Assign the two disks with some data as disk1 and disk2. Since you have no valid parity, you can set a new disk configuration. Remember, do NOT assign a parity disk. Then, you'll have two disks that will not mount and both will show as unformatted. DO NOT FORMAT THEM!!!!!!! Run reiserfsck --rebuild-sb /dev/md1 and reiserfsck --rebuild-sb /dev/md2 remember, the correct answers to the rebuild-sb prompts were in the post I linked to in my prior answer. Joe L. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Ok. So I'll do a --rebuild-sb on the 2 2TB disks. How do I get rid of the invalid configuration, because I need to have /dev/mdX available. I presume the array has to be down Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 You'll need to set a new initial configuration to bring the array online so /dev/md1 and /dev/md2 are available. Just do not connect the third disk as of yet. You can then try recovering data from both disks by rebuilding the superblock on both. One will do better (hopefully) When the reiserfsck finishes, it will suggest what to do next. Joe L. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 Hi, Did a rebuild-sd and a check. Check failed on the disk I think is the parity disk Afterwards I stopped the array and reboot. Syslog attached. Drives still are shown as unformatted. syslog.txt Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 Hi, Did a rebuild-sd and a check. Check failed on the disk I think is the parity disk Afterwards I stopped the array and reboot. Syslog attached. Drives still are shown as unformatted. Both disks still have issues, but I expected that. I would now try reiserfsck --fix-fixable /dev/md1 and reiserfsck --fix-fixable /dev/md2 Edit: fixed typo Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 I attached the screenshot and syslog Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 and the syslog and md2 log syslog.txt Quote Link to comment
Joe L. Posted March 27, 2013 Share Posted March 27, 2013 They both need the same rebuild of their internal file-tree. reiserfsck --rebuild-tree /dev/md1 reiserfsck --rebuild-tree /dev/md2 This step may take a while if the disk was full. Joe L. Quote Link to comment
deva1983 Posted March 27, 2013 Author Share Posted March 27, 2013 (some) data loss is to be expected ? 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.