April 29, 201412 yr jonathanm, Thank you. That's quite interesting. I am not doing anything to the browser, yet somehow it keeps acting like that. When I started the server without the plugins it shut itself down beautifully. Dynamix is the culprit.
April 29, 201412 yr The only plugins I am running are:BTW, it's a very bad idea to be running ANY plugins when your array isn't healthy. I don't think any of the plugin authors test their software additions to unraid with the array in a bad situation, they all assume everything is working as it should. Trying to solve an array issue with plugins running is going to complicate matters immensely.
April 29, 201412 yr Author @jonathanm, Thank you! Yes, I agree wholeheartedly that running multiple plugins are not good when the array is in jeopardy. This is the reason that I rebooted the server in safe-mode and stopped the array. I'm counting on bjp999 to help me through the reconstruction process (and whatever can be salvaged) so I'm keeping everything off until I can get the next step from him. Thanks again for your help! ps. Out of curiosity, how did you know that Dynamix was causing the problem with the shutdown? I will definitely get rid of dynamix, because other than being eye-candy it doesnt serve much purpose for me.
April 30, 201412 yr Author Wow... Your life is as bad as mine! Leaving work a 9pm! No rush, if you can't get to it. I can look for your answer tomorrow. This is voluntary work that you are doing by helping a kindred soul so I appreciate every bit of it, but not at the expense of your personal sanity! Thank you again!
April 30, 201412 yr Step 1 - Mount disk3 into the cache slot: Before we can just put the drive in the cache slot we need to first unassign it from its current slot. So you need to stop the array, go to slot #3 (red balled slot), unassign the disk. I believe that is your Samsung drive with serial number ending with 3152. Start the array. (Drive 3 should say something like missing or unassigned). Then stop the array again. Make sure that unRAID doesn't automatically reassign a disk to slot 3 again. If it does just unassign it. Then assign the disk (Samsung ending with 3152) that was in the disk3 slot to the cache slot. Start the array. This is not something I do every day, so I can't tell you exactly what you will see. If in doubt take a screenshot and post it. You should then, from your windows workstation, be able to go to \\tower\cache share. Please do that and have a look around. This should be the old disk3. See if the files are valid and accurate. Hopefully disk3 will look good. DO NOT WRITE TO THE DISK!!!! This in incredibly important - no writes to any disks in the array. Do not try to play a movie from a media player that is going to create a file on the disk (as they like to do to remember the current location). If necessary copy a movie to your workstation and try to play it from there. Convince yourself that disk3 is good, then post back.
April 30, 201412 yr ps. Out of curiosity, how did you know that Dynamix was causing the problem with the shutdown? I will definitely get rid of dynamix, because other than being eye-candy it doesnt serve much purpose for me. I don't know that it was causing the shutdown issue, but it definitely was causing the web gui reload syslog spam. Any program that keeps a file open on the array will cause a shutdown issue. Theoretically plugins should monitor requests to stop the array and nicely terminate so the array can shut down, but that doesn't always work well when the array is healthy, so who knows how they will react when the array is having issues. I wouldn't necessarily throw dynamix under the bus, but you need to read the entire dynamix thread to understand the progression of things. I'm pretty sure the constant reload issue was discussed in the main thread, there is a setting to revert to manual refresh like the stock gui.
April 30, 201412 yr ps. Out of curiosity, how did you know that Dynamix was causing the problem with the shutdown? I will definitely get rid of dynamix, because other than being eye-candy it doesnt serve much purpose for me. I don't know that it was causing the shutdown issue, but it definitely was causing the web gui reload syslog spam. Any program that keeps a file open on the array will cause a shutdown issue. Theoretically plugins should monitor requests to stop the array and nicely terminate so the array can shut down, but that doesn't always work well when the array is healthy, so who knows how they will react when the array is having issues. I wouldn't necessarily throw dynamix under the bus, but you need to read the entire dynamix thread to understand the progression of things. I'm pretty sure the constant reload issue was discussed in the main thread, there is a setting to revert to manual refresh like the stock gui. +1 Dynamix (by default) continually updates page status...which is usually the behavior people want. That can be turned off in settings.
April 30, 201412 yr +1 Dynamix (by default) continually updates page status...which is usually the behavior people want. That can be turned off in settings. Problem is that it spams the syslog. Dynamix is working at odds with the design of the stock gui to achieve something that should be handled differently. Normally it doesn't cause too much of an issue, except when you run out of space for the syslog, which causes troubleshooting nightmares. It would be nice if the default was NOT to continually update, with a warning notice about the syslog messages if you enable the auto refresh option.
April 30, 201412 yr Step 1 - Mount disk3 into the cache slot: Before we can just put the drive in the cache slot we need to first unassign it from its current slot. So you need to stop the array, go to slot #3 (red balled slot), unassign the disk. I believe that is your Samsung drive with serial number ending with 3152. Start the array. (Drive 3 should say something like missing or unassigned). Then stop the array again. Make sure that unRAID doesn't automatically reassign a disk to slot 3 again. If it does just unassign it. Then assign the disk (Samsung ending with 3152) that was in the disk3 slot to the cache slot. Start the array. This is not something I do every day, so I can't tell you exactly what you will see. If in doubt take a screenshot and post it. You should then, from your windows workstation, be able to go to \\tower\cache share. Please do that and have a look around. This should be the old disk3. See if the files are valid and accurate. Hopefully disk3 will look good. DO NOT WRITE TO THE DISK!!!! This in incredibly important - no writes to any disks in the array. Do not try to play a movie from a media player that is going to create a file on the disk (as they like to do to remember the current location). If necessary copy a movie to your workstation and try to play it from there. Convince yourself that disk3 is good, then post back. Success?
April 30, 201412 yr Author The process went well. As you said, the server unassigned the SAMSUNG drive and reassigned it to the cache drive (and the disk3 slot remained 'unassigned'). The array was started. I was able to connect to the cache share and browse through the contents of disk3/SAMSUNG drive. As a test, I tried to copy a 1 GB movie file from the drive into my local hard drive. The copy process progressed to 99% and then get stuck and I got an error saying 'cannot establish network connection'. There is clearly nothing wrong with the network (I tested and verified), so that leaves me to believe that this is either a filesystem error that is preventing the copy or a physical problem with the SAMSUNG drive. Other than that, I can telnet into my unraid box, and browse through all the way to individual file levels on the SAMSUNG drive. Every single directory was tested and I had no problem browsing. Will await your instructions... (p.s. I am working tonight overnight and won't get home until tomorrow so will post the results of the next step when i get back). Thanks so much for your help!
April 30, 201412 yr The process went well. As you said, the server unassigned the SAMSUNG drive and reassigned it to the cache drive (and the disk3 slot remained 'unassigned'). The array was started. I was able to connect to the cache share and browse through the contents of disk3/SAMSUNG drive. As a test, I tried to copy a 1 GB movie file from the drive into my local hard drive. The copy process progressed to 99% and then get stuck and I got an error saying 'cannot establish network connection'. There is clearly nothing wrong with the network (I tested and verified), so that leaves me to believe that this is either a filesystem error that is preventing the copy or a physical problem with the SAMSUNG drive. Other than that, I can telnet into my unraid box, and browse through all the way to individual file levels on the SAMSUNG drive. Every single directory was tested and I had no problem browsing. Will await your instructions... (p.s. I am working tonight overnight and won't get home until tomorrow so will post the results of the next step when i get back). Thanks so much for your help! Not good enough that you can browse the directories. I need to know if the data is good. I would not interpret a copy ending at 99% as a file system error. Post a syslog. Keep trying. Different files. Check at least 10 files from different locations. Try to check a variety of files - some copied when you first set up the array and others more recently copied. When you browsed did you browse /mnt/cache?
April 30, 201412 yr Author OK, so I think the data is good (for the most part) -- I say that since I'm not sure whether there are areas I haven't discovered. So far, I copies approximately 10-12 different files (ranging from photos, movies, tv shows) over to my local computer and tested them and they all look good. In fact, I was also able to copy the file that I previously couldn't (due to a network error). I went through /mnt/cache/movies; /mnt/cache/tvshows; /mnt/cache/music; mnt/cache/photos. All of the directories are browseable. All the data that I transferred so far play well on my local computer. As you have said, I have not written a thing to the drive (disk3). Every test was run by copying the file over to a local computer and then running it off the local computer. I even tested them using a PC (I am a mac user) since I am worried about the .ds_store and all the other garbage files from the Apple finder being generated on the cache drive. The files I tested ranged from a few KB to several GB. In fact, I successfully copied ALL of the family photos, family videos that were stored on that disk on the array. I am now trying to see if I could copy all of the movies off that drive (I have about 200 GB worth) on to my desktop. I will report back later with the results. I also wonder -- could the errors be due to a defective backplane on my Norco 4220 chassis, or due to a defective SATA controller? When the disk3 and disk4 were giving me errors, both drives were run on two different backplanes on my chassis, with the same SATA controller card (SuperMicro AOC-SAS2LP-MV8), but now the "defective" drive is connected to a direct SATA port on the motherboard. Anyways, I'm curious as to what you think I should do next.
April 30, 201412 yr The next step is to reconstruct the array. I will post instructions tonight or tomorrow morning.
April 30, 201412 yr Actually I am a little shakey on the exact steps. I will lay them out generally and ask others to help with the exact steps. So don't do this until the steps are clean and agreed ... 1. Shut down array 2. Note drive locations (which drives are assigned to which slots) 3. Do a new config. Somehow you can specify that to trust parity and not to rebuild it. I have not done this with the 5.0.5 version, so someone will need to fill in the blanks. 4. Assign each drive to its slot (including disk4) 5. Start the array. Should be a non-event. 6. Stop the array. 7. Unassign disk4. 8. Start the array. 9. Disk4 will now be simulated. You should be able to look at /mnt/disk4 and see valid contents. This is what will be constructed. You can try to access from a workstation to confirm the files are good. Although I wouldn’t advise not to write to it, the truth is it won’t hurt anything at this point. 10. Stop the array 11. Assign the real disk to the disk4 slot again 12. When you go to start the array it should say something about starting the array and rebuilding the disk. Again, maybe someone can confirm exactly what you should see. 13. Start the array. The disk will rebuild 14. After it is complete, disk4 will be rebuilt. It should look exactly like the simulated disk4 you saw in step 9. Hope no other disks fail during the rebuild!! I’m hopeful the results will be better. Know that any writes to any disks outside the array will create small corruptions. I’m hopeful these will be few and far between. 15. Done
April 30, 201412 yr Actually I am a little shakey on the exact steps. I will lay them out generally and ask others to help with the exact steps. So don't do this until the steps are clean and agreed ... 1. Shut down array 2. Note drive locations (which drives are assigned to which slots) 3. Do a new config. Somehow you can specify that to trust parity and not to rebuild it. I have not done this with the 5.0.5 version, so someone will need to fill in the blanks. 4. Assign each drive to its slot (including disk4) 5. Start the array. Should be a non-event. 6. Stop the array. 7. Unassign disk4. 8. Start the array. 9. Disk4 will now be simulated. You should be able to look at /mnt/disk4 and see valid contents. This is what will be constructed. You can try to access from a workstation to confirm the files are good. Although I wouldn’t advise not to write to it, the truth is it won’t hurt anything at this point. 10. Stop the array 11. Assign the real disk to the disk4 slot again 12. When you go to start the array it should say something about starting the array and rebuilding the disk. Again, maybe someone can confirm exactly what you should see. 13. Start the array. The disk will rebuild 14. After it is complete, disk4 will be rebuilt. It should look exactly like the simulated disk4 you saw in step 9. Hope no other disks fail during the rebuild!! I’m hopeful the results will be better. Know that any writes to any disks outside the array will create small corruptions. I’m hopeful these will be few and far between. 15. Done This looks good. There is a checkbox for step 3.
April 30, 201412 yr Author OK, this sounds good. But is it not necessary to do a file system check for the disks before I try to rebuild them? This is the sequence of steps that Garycase suggested to me some time ago (I did this exact sequence of steps) and it led to the failure of rebuild of disk4 because disk3 was acting like it was corrupted during the rebuild. Anyways, I've nearly copied 90% of my movie collection off disk3 to my local hard drive, so I am not that hard broken about the loss of data on that drive. I hope this works. I will give it a shot tomorrow and report back.
May 1, 201412 yr Author OK, I've run into a snag. I'm wondering whether this means that my disk4 data is lost. I somehow managed to backup the data for disk3 and made a listing of disk4 (i.e. the movies and tv shows I need) before it went blast. Here's the problem I run into. 1. Shut down array ? Done. 2. Note drive locations (which drives are assigned to which slots) ? Done 3. Do a new config (with parity trust option) ? Done 4. Assign each drive to its slot (including disk4) ? Done 5. Start the array. ? Done. But when I do so, the array starts normally, but disk4 shows up as "unformatted" see screenshot below. 6. Stop the array ? Done 7. Unassign disk4. ? Done. 8. Start the array ? Done. Disk4 still shows up as "unformatted" (see screenshot below) 9. Disk4 will now be simulated. You should be able to look at /mnt/disk4 and see valid contents. This is what will be constructed. You can try to access from a workstation to confirm the files are good. Although I wouldn’t advise not to write to it, the truth is it won’t hurt anything at this point. ? Here is the problem. Disk4 is actually NOT simulated. I have my /mnt/user/* shares, but the /mnt/disk4 directory is altogether missing. disk1, disk2, disk3, disk5, disk6 are all present. Moreover, when I browse over to, say, /mnt/user/movies, the movies that were previously on disk4 are not there (I suppose this is to be expected since unraid is not simulating disk4) . 10. Stop the array ? Done. 11. Assign the real disk to the disk4 slot again ? Done. 12. When you go to start the array it should say something about starting the array and rebuilding the disk. ? Done 13. Start the array. The disk will rebuild ? Yes, the rebuild process automatically started. But since I did not get simulated disk4, I stopped the rebuild process and went back to scratch until I could hear back from you if there is an interim step that needed to be done. What do you think? I fear this may mean that I may not get my disk4 contents back. Should I just proceed with the rebuild process?
May 1, 201412 yr You may be right. Disk4 may be lost. I would suggest, though, that you allow the rebuild of disk4 to complete. Then we can try to use Reiserfsck to do a full recovery on the reconstructed disk. I did ask that you confirm at step 9 that the disk looked good. I don't know why you continued past that step...
May 1, 201412 yr Author The data rebuild is in progress. But disk4 is still showing up as "unformatted". I captured the syslog during the time that the formatting "appeared" to start before it started rebuilding. Is this a problem? There are some errors about a "failed mount" that I noted. May 1 10:18:03 Tower emhttp_event: started May 1 10:18:04 Tower avahi-daemon[7834]: Service "Tower" (/services/smb.service) successfully established. May 1 10:18:19 Tower emhttp: shcmd (1369): set -o pipefail ; mkreiserfs -q /dev/md4 |& logger May 1 10:18:19 Tower logger: mkreiserfs 3.6.24 May 1 10:18:19 Tower logger: May 1 10:22:27 Tower emhttp: shcmd (1370): mkdir /mnt/disk4 May 1 10:22:27 Tower emhttp: shcmd (1371): set -o pipefail ; mount -t reiserfs -o user_xattr,acl,noatime,nodiratime /dev/md4 /mnt/disk4 |& logger May 1 10:22:27 Tower kernel: REISERFS (device md4): found reiserfs format "3.6" with standard journal May 1 10:22:27 Tower kernel: REISERFS (device md4): using ordered data mode May 1 10:22:27 Tower kernel: reiserfs: using flush barriers May 1 10:22:27 Tower kernel: REISERFS (device md4): journal params: device md4, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 May 1 10:22:27 Tower kernel: REISERFS (device md4): checking transaction log (md4) May 1 10:23:49 Tower kernel: REISERFS (device md4): Using r5 hash to sort names May 1 10:23:50 Tower logger: mount: Operation not supported May 1 10:23:50 Tower emhttp: _shcmd: shcmd (1371): exit status: 32 May 1 10:23:50 Tower emhttp: disk4 mount error: 32 May 1 10:23:50 Tower emhttp: shcmd (1372): rmdir /mnt/disk4 May 1 10:23:50 Tower kernel: ------------[ cut here ]------------ May 1 10:23:50 Tower kernel: WARNING: at fs/inode.c:974 unlock_new_inode+0x23/0x40() May 1 10:23:50 Tower kernel: Hardware name: X10SL7-F May 1 10:23:50 Tower kernel: Modules linked in: md_mod sg coretemp mvsas libsas acpi_cpufreq i2c_i801 ahci igb libahci hwmon i2c_algo_bit i2c_core ptp pps_core mpt2sas scsi_transport_sas raid_class mperf [last unloaded: md_mod] May 1 10:23:50 Tower kernel: Pid: 8055, comm: mount Not tainted 3.9.11p-unRAID #5 May 1 10:23:50 Tower kernel: Call Trace: May 1 10:23:50 Tower kernel: [<c1029269>] warn_slowpath_common+0x77/0x8e May 1 10:23:50 Tower kernel: [<c10a78f1>] ? unlock_new_inode+0x23/0x40 May 1 10:23:50 Tower kernel: [<c10a78f1>] ? unlock_new_inode+0x23/0x40 May 1 10:23:50 Tower kernel: [<c102929d>] warn_slowpath_null+0x1d/0x1f May 1 10:23:50 Tower kernel: [<c10a78f1>] unlock_new_inode+0x23/0x40 May 1 10:23:50 Tower kernel: [<c10e526f>] reiserfs_new_inode+0x599/0x5ad May 1 10:23:50 Tower kernel: [<c10fbd7a>] ? reiserfs_get_acl+0xcf/0x1c5 May 1 10:23:50 Tower kernel: [<c10e018f>] reiserfs_mkdir+0x138/0x236 May 1 10:23:50 Tower kernel: [<c10fa27a>] reiserfs_xattr_init+0xde/0x201 May 1 10:23:50 Tower kernel: [<c10eb55e>] reiserfs_fill_super+0x75e/0x8a7 May 1 10:23:50 Tower kernel: [<c1597000>] ? pcibios_setup+0xaa/0x38e May 1 10:23:50 Tower kernel: [<c1098d6f>] mount_bdev+0x107/0x150 May 1 10:23:50 Tower kernel: [<c107e956>] ? pcpu_alloc+0x261/0x2e0 May 1 10:23:50 Tower kernel: [<c10e9464>] get_super_block+0x1a/0x1f May 1 10:23:50 Tower kernel: [<c10eae00>] ? read_super_block+0x284/0x284 May 1 10:23:50 Tower kernel: [<c1097e0d>] mount_fs+0x16/0xae May 1 10:23:50 Tower kernel: [<c107e9ee>] ? __alloc_percpu+0xa/0xc May 1 10:23:50 Tower kernel: [<c10ab196>] ? alloc_vfsmnt+0x97/0x112 May 1 10:23:50 Tower kernel: [<c10ac995>] vfs_kern_mount+0x51/0xa9 May 1 10:23:50 Tower kernel: [<c10aca97>] do_new_mount+0x85/0x129 May 1 10:23:50 Tower kernel: [<c10acc49>] do_mount+0x10e/0x19d May 1 10:23:50 Tower kernel: [<c10acd40>] sys_mount+0x68/0x94 May 1 10:23:50 Tower kernel: [<c1401190>] syscall_call+0x7/0xb May 1 10:23:50 Tower kernel: ---[ end trace 850111fb4755c09b ]--- May 1 10:23:50 Tower kernel: REISERFS warning (device md4): jdm-20006 create_privroot: xattrs/ACLs enabled and couldn't find/create .reiserfs_priv. Failing mount. May 1 10:23:51 Tower emhttp: shcmd (1373): :>/etc/samba/smb-shares.conf May 1 10:23:51 Tower emhttp: shcmd (1374): cp /etc/exports- /etc/exports May 1 10:23:51 Tower avahi-daemon[7834]: Files changed, reloading. May 1 10:23:51 Tower emhttp: shcmd (1375): echo '"/mnt/user/bollywood" -async,no_subtree_check,fsid=108 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1376): echo '"/mnt/user/family_videos" -async,no_subtree_check,fsid=107 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: get_config_idx: fopen /boot/config/shares/lost+found.cfg: No such file or directory - assigning defaults May 1 10:23:51 Tower emhttp: shcmd (1377): echo '"/mnt/user/movies" -async,no_subtree_check,fsid=100 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1378): echo '"/mnt/user/music" -async,no_subtree_check,fsid=106 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1379): echo '"/mnt/user/other" -async,no_subtree_check,fsid=105 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1380): echo '"/mnt/user/photos" -async,no_subtree_check,fsid=102 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1381): echo '"/mnt/user/sinhala" -async,no_subtree_check,fsid=104 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1382): echo '"/mnt/user/tvshows" -async,no_subtree_check,fsid=101 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: shcmd (1383): echo '"/mnt/user/unsorted" -async,no_subtree_check,fsid=103 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports May 1 10:23:51 Tower emhttp: Restart SMB... May 1 10:23:51 Tower emhttp: shcmd (1384): killall -HUP smbd May 1 10:23:51 Tower emhttp: shcmd (1385): cp /etc/avahi/services/smb.service- /etc/avahi/services/smb.service May 1 10:23:51 Tower avahi-daemon[7834]: Files changed, reloading. May 1 10:23:51 Tower avahi-daemon[7834]: Service group file /services/smb.service changed, reloading. May 1 10:23:51 Tower emhttp: shcmd (1386): ps axc | grep -q rpc.mountd May 1 10:23:51 Tower emhttp: Restart NFS... May 1 10:23:51 Tower emhttp: shcmd (1387): exportfs -ra |& logger May 1 10:23:51 Tower emhttp: shcmd (1388): /usr/local/sbin/emhttp_event svcs_restarted May 1 10:23:51 Tower emhttp_event: svcs_restarted May 1 10:23:52 Tower avahi-daemon[7834]: Service "Tower" (/services/smb.service) successfully established.
May 2, 201412 yr Author So I am done with the rebuild process. Unlike last time, this time, I haven't gotten any errors that accumulated in the system. However, disk4 is still showing up as "unformatted" after the data rebuild. See the screenshot below. I browsed through the directories using telnet, but while I see /mnt/disk1, /mnt/disk2, mnt/disk3, I do not see mnt/disk4. (disk5 and disk6 are also present). When I browse through the user share, (/mnt/user), the files that would have been on disk4 are missing. Lastly, when I tried : df -h, I get the following output: root@Tower:/mnt/user# df -h Filesystem Size Used Avail Use% Mounted on tmpfs 128M 2.4M 126M 2% /var/log /dev/sda1 7.5G 94M 7.4G 2% /boot /dev/md1 1.9T 1.9T 0 100% /mnt/disk1 /dev/md2 1.9T 1.8T 123G 94% /mnt/disk2 /dev/md3 1.9T 1.6T 299G 84% /mnt/disk3 /dev/md5 3.7T 2.7T 1.1T 73% /mnt/disk5 /dev/md6 2.8T 2.6T 216G 93% /mnt/disk6 shfs 12T 11T 1.7T 87% /mnt/user What should I try at this point? Should I try formatting pre-clearing disk4 and re-trying, or should I just try to use reiserfsck to fix the file structure? I will wait till I hear back from you.
May 2, 201412 yr Stop array. Start array in maintenance mode Run the command reiserfsck --rebuild-sb /dev/md4 It will ask you a bunch of questions. See THIS THREAD for the answers. When this command completes you will then need to run this command reiserfsck --rebuild-tree --scan-whole-partition /dev/md4 It will run for hours. When it is done, you need to stop the array, and then start the array. Disk4 should be better. There may be a bunch of junky files and empty directories, but if you keep looking I;m hopeful you will find some of your data. Good luck. (If at any point you are unsure what to do, post back)
May 2, 201412 yr Author OK, I started the array in maintenance mode and tried the reiserfsck --rebuild-sb /dev/md4 The following is the output I got: root@Tower:~# reiserfsck --rebuild-sb /dev/md4 reiserfsck 3.6.24 Will check superblock and rebuild it if needed Will put log info to 'stdout' Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes Reiserfs super block in block 16 on 0x904 of format 3.6 with standard journal Count of blocks on the device: 976754624 Number of bitmaps: 29809 Blocksize: 4096 Free blocks (count of blocks - used [journal, bitmaps, data, reserved] blocks): 976716604 Root block: 8211 Filesystem is clean Tree height: 2 Hash function used to sort names: "r5" Objectid map size 2, max 972 Journal parameters: Device [0x0] Magic [0x65072f9a] Size 8193 blocks (including 1 for journal header) (first block 18) Max transaction length 1024 blocks Max batch size 900 blocks Max commit age 30 Blocks reserved by journal: 0 Fs state field: 0x0: sb_version: 2 inode generation number: 0 UUID: 9c826b5b-a35f-4310-97d3-9a6ceed25118 LABEL: Set flags in SB: ATTRIBUTES CLEAN Mount count: 1 Maximum mount count: 30 Last fsck run: Thu May 1 10:18:19 2014 Check interval in days: 180 Super block seems to be correct Looks like it is saying that the superblock is OK. I then did the reiserfsck --rebuild-tree --scan-whole-partition /dev/md4 root@Tower:~# reiserfsck --rebuild-tree --scan-whole-partition /dev/md4 reiserfsck 3.6.24 ************************************************************* ** Do not run the program with --rebuild-tree unless ** ** something is broken and MAKE A BACKUP before using it. ** ** If you have bad sectors on a drive it is usually a bad ** ** idea to continue using it. Then you probably should get ** ** a working hard drive, copy the file system from the bad ** ** drive to the good one -- dd_rescue is a good tool for ** ** that -- and only then run this program. ** ************************************************************* Will rebuild the filesystem (/dev/md4) tree Will put log info to 'stdout' Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes Replaying journal: No transactions found .... and I haven't seen any output since then. It's been going for a while now. I will check back in the morning and see if there is progression.
May 2, 201412 yr OK, I started the array in maintenance mode and tried the reiserfsck --rebuild-sb /dev/md4 The following is the output I got: root@Tower:~# reiserfsck --rebuild-sb /dev/md4 reiserfsck 3.6.24 Will check superblock and rebuild it if needed Will put log info to 'stdout' Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes Reiserfs super block in block 16 on 0x904 of format 3.6 with standard journal Count of blocks on the device: 976754624 Number of bitmaps: 29809 Blocksize: 4096 Free blocks (count of blocks - used [journal, bitmaps, data, reserved] blocks): 976716604 Root block: 8211 Filesystem is clean Tree height: 2 Hash function used to sort names: "r5" Objectid map size 2, max 972 Journal parameters: Device [0x0] Magic [0x65072f9a] Size 8193 blocks (including 1 for journal header) (first block 18) Max transaction length 1024 blocks Max batch size 900 blocks Max commit age 30 Blocks reserved by journal: 0 Fs state field: 0x0: sb_version: 2 inode generation number: 0 UUID: 9c826b5b-a35f-4310-97d3-9a6ceed25118 LABEL: Set flags in SB: ATTRIBUTES CLEAN Mount count: 1 Maximum mount count: 30 Last fsck run: Thu May 1 10:18:19 2014 Check interval in days: 180 Super block seems to be correct Looks like it is saying that the superblock is OK. I then did the reiserfsck --rebuild-tree --scan-whole-partition /dev/md4 root@Tower:~# reiserfsck --rebuild-tree --scan-whole-partition /dev/md4 reiserfsck 3.6.24 ************************************************************* ** Do not run the program with --rebuild-tree unless ** ** something is broken and MAKE A BACKUP before using it. ** ** If you have bad sectors on a drive it is usually a bad ** ** idea to continue using it. Then you probably should get ** ** a working hard drive, copy the file system from the bad ** ** drive to the good one -- dd_rescue is a good tool for ** ** that -- and only then run this program. ** ************************************************************* Will rebuild the filesystem (/dev/md4) tree Will put log info to 'stdout' Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes Replaying journal: No transactions found .... and I haven't seen any output since then. It's been going for a while now. I will check back in the morning and see if there is progression. Its good news there is a superblock I think. It may or may not be done in the morning. Whatever you do do not stop it. Let it go several days if necessary!
Archived
This topic is now archived and is closed to further replies.