February 17, 201115 yr Author Yes, but now if I re-assign the parity disk, unraid will first start a sync. How can I start it and let unraid know the parity disk is still valid, so I can format the md9 disk with protection enabled? Anyway, I will made a final parity check at the end of this process, just to be sure it was an option.
February 17, 201115 yr Yes, but now if I re-assign the parity disk, unraid will first start a sync. How can I start it and let unraid know the parity disk is still valid, so I can format the md9 disk with protection enabled? Anyway, I will made a final parity check at the end of this process, just to be sure it was an option. You can use a modified version of the trust-my-parity process. By using the /root/mdcmd set invalidslot NN command you can force the server to consider any "slot" in the array to be invalid, or, specify an on-existent disk to have them all be considered valid. So, if you want to force disk9 to be considered invalid and to be rebuilt, you would Stop the array Log in via telnet Type initconfig respond to its prompt with Yes (Capital "Y", lower case "es") Then, before starting the array type /root/mdcmd set invalidslot 9 It will force the array to think slot 9 is the one to be re-constructed. When you then start the array it will re-construct disk9 from what it is reading from parity in combination with the other drives. Joe L.
February 17, 201115 yr Author Thanks for your answer, even, It's absolutely not what I want to do. I don't want to reconstruct the disk 9 with bad data written in the parity disk by the parity check done when I start the array just after the MBR write on the disk 9. I've made a copy of the data still valid from disk 9 to the new disk, so I don't need the data from the disk 9 or from the parity anymore (except if another disk fail now). Now I really want to clean the disk 9, there are errors on it, I want to correct it, actually. And if possible, with protection. The disk 9 is unmounted ("unformatted"), the parity disk is disabled, and I really don't know what to do.
February 17, 201115 yr Thanks for your answer, even, It's absolutely not what I want to do. I don't want to reconstruct the disk 9 with bad data written in the parity disk by the parity check done when I start the array just after the MBR write on the disk 9. I've made a copy of the data still valid from disk 9 to the new disk, so I don't need the data from the disk 9 or from the parity anymore (except if another disk fail now). Now I really want to clean the disk 9, there are errors on it, I want to correct it, actually. And if possible, with protection. The disk 9 is unmounted ("unformatted"), the parity disk is disabled, and I really don't know what to do. Sorry. I mis-understood your request. If there is nothing on disk9 you wish, and if parity is disabled, then you can just re-format it. The command for that is mkreiserfs /dev/md9 If you want to correct the errors, and possibly get to files it can recover to the lost+found directory it will create, and not just re-format it, you can type reiserfsck --check /dev/md9 then follow its suggestions. for which option to use next. Most times it will suggest --fix-fixable first. When you eventually re-assign parity unRAID will completely rebuild parity based on the then assigned and working disks. Joe L.
February 17, 201115 yr Sorry coming late to the game, but if parity is not valid, would it be possible to use the "/dev/sdX" rather than "/dev/mdX" device, and speed up the reiserfsck?
February 17, 201115 yr Sorry coming late to the game, but if parity is not valid, would it be possible to use the "/dev/sdX" rather than "/dev/mdX" device, and speed up the reiserfsck? Yes, but then there is the confusion that always seems to occur with users running the file system check on /dev/sdX instead of /dev/sdX1 and making things worse when the superblock is not found. For that reason /dev/mdX is better. Joe L.
February 17, 201115 yr Author The disk 9 is unmounted ("unformatted"), the parity disk is disabled, and I really don't know what to do. Sorry. I mis-understood your request. If there is nothing on disk9 you wish, and if parity is disabled, then you can just re-format it. The command for that is mkreiserfs /dev/md9 root@Tower:~# mkreiserfs /dev/md9 mkreiserfs 3.6.21 (2009 www.namesys.com) A pair of credits: Oleg Drokin was the debugger for V3 during most of the time that V4 was under development, and was quite skilled and fast at it. He wrote the large write optimization of V3. Elena Gryaznova performed testing and benchmarking. Guessing about desired format.. Kernel 2.6.32.9-unRAID is running. Format 3.6 with standard journal Count of blocks on the device: 244190624 Number of blocks consumed by mkreiserfs formatting process: 15664 Blocksize: 4096 Hash function used to sort names: "r5" Journal Size 8193 blocks (first block 18) Journal Max transaction length 1024 inode generation number: 0 UUID: 0c998131-fe38-4db4-8cba-5e8364f4d71c ATTENTION: YOU SHOULD REBOOT AFTER FDISK! ALL DATA WILL BE LOST ON '/dev/md9'! Continue (y/n):y Initializing journal - 0%....20%....40%....60%....80%....100% Syncing..ok ReiserFS is successfully created on /dev/md9. root@Tower:~# Now on the main windows disk9 is still unformatted, after refresh. I suppose I have to Stop...Reboot...re-assign parity disk..."I'm sure..."..Start... and wait till done. Right?
February 17, 201115 yr Stop and reboot, it should show that disk9 has the entire disk free. Then you can stop the array and re-assign the parity drive. When you next start the array parity will be calculated. Joe L.
February 18, 201115 yr Author Parity Sync reconstruction, around 900 minutes left. I have peek at the syslog, and there was a problem (red in syslog) yesterday. I've launch the rebuilding at around 16h30, and the problem appears at around 23h15. Feb 17 16:28:44 Tower emhttp: shcmd (152): rm /etc/samba/smb-shares.conf >/dev/null 2>&1 Feb 17 16:28:44 Tower emhttp: shcmd (153): cp /etc/exports- /etc/exports Feb 17 16:28:44 Tower emhttp: shcmd (154): mkdir /mnt/user Feb 17 16:28:44 Tower emhttp: shcmd (155): /usr/local/sbin/shfs /mnt/user -o noatime,big_writes,allow_other,default_permissions Feb 17 16:28:45 Tower emhttp: shcmd (156): killall -HUP smbd Feb 17 16:28:45 Tower emhttp: shcmd (157): /etc/rc.d/rc.nfsd restart | logger Feb 17 23:15:19 Tower kernel: ata5: lost interrupt (Status 0x51) Feb 17 23:15:19 Tower kernel: ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x280000 action 0x6 frozen Feb 17 23:15:19 Tower kernel: ata5: SError: { 10B8B BadCRC } Feb 17 23:15:19 Tower kernel: ata5.00: failed command: READ DMA EXT Feb 17 23:15:19 Tower kernel: ata5.00: cmd 25/00:00:97:20:eb/00:04:1f:00:00/e0 tag 0 dma 524288 in Feb 17 23:15:19 Tower kernel: res 40/00:ff:00:00:00/00:00:00:00:00/50 Emask 0x4 (timeout) Feb 17 23:15:19 Tower kernel: ata5.00: status: { DRDY } Feb 17 23:15:19 Tower kernel: ata5: hard resetting link Feb 17 23:15:20 Tower kernel: ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Feb 17 23:15:20 Tower kernel: ata5.00: configured for UDMA/100 Feb 17 23:15:20 Tower kernel: ata5.00: device reported invalid CHS sector 0 Feb 17 23:15:20 Tower kernel: ata5: EH complete Feb 18 15:00:13 Tower unmenu[1711]: gawk: ./08-unmenu-array_mgmt.awk:115: warning: escape sequence `\'' treated as plain `'' Of cours, I've absolutely no idea about what it's speaking about : ata5? Everything is in sata, the usb maybe? Is it important? If yes, is there a cure? Thanx again
Archived
This topic is now archived and is closed to further replies.