Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Fri, Jun 18, 2010 at 9:11 AM, Huang, Tao deb...@huangtao.me wrote: On Fri, Jun 18, 2010 at 6:02 PM, Tom H tomh0...@gmail.com wrote: [snip] mdadm assembles an array according to data in the superblock so it shouldn't matter whether the kernel recognizes sda and sdb as sdb and sda respectively should you plug them in differently. so they's recognized with data in the superblock, that even uuid doesn't matter? UUIDs are held in superblocks. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimnsbxvxjuu4lerzkledis0oe470dtpu4y2y...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On 17/06/2010 14:08, Huang, Tao wrote: On Thu, Jun 17, 2010 at 4:17 PM, Michal mic...@ionic.co.uk wrote: This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) I think the allocating of /dev/sdX depends on the order you plug the drives into the machine. so it changes over reconfiguring of the hardwares, which makes your labels useless. can someone confirm this? Tao But how can this be correct when each raid partion is linked to the HDD/Partions # cat /proc/mdstat Personalities : [raid1] [raid6] [raid5] [raid4] md3 : active raid1 sda2[0] sdb2[1] 716796096 blocks [2/2] [UU] md2 : active raid1 sda5[0] sdb5[1] 51199040 blocks [2/2] [UU] md0 : active raid1 sda1[0] sdb1[1] 513984 blocks [2/2] [UU] md1 : active raid1 sda3[0] sdb3[1] 102398208 blocks [2/2] [UU] unused devices: none for example? -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c1b2e06.5060...@ionic.co.uk
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Fri, Jun 18, 2010 at 4:27 AM, Michal mic...@ionic.co.uk wrote: On 17/06/2010 14:08, Huang, Tao wrote: On Thu, Jun 17, 2010 at 4:17 PM, Michal mic...@ionic.co.uk wrote: This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) I think the allocating of /dev/sdX depends on the order you plug the drives into the machine. so it changes over reconfiguring of the hardwares, which makes your labels useless. can someone confirm this? Tao But how can this be correct when each raid partion is linked to the HDD/Partions # cat /proc/mdstat Personalities : [raid1] [raid6] [raid5] [raid4] md3 : active raid1 sda2[0] sdb2[1] 716796096 blocks [2/2] [UU] md2 : active raid1 sda5[0] sdb5[1] 51199040 blocks [2/2] [UU] md0 : active raid1 sda1[0] sdb1[1] 513984 blocks [2/2] [UU] md1 : active raid1 sda3[0] sdb3[1] 102398208 blocks [2/2] [UU] mdadm assembles an array according to data in the superblock so it shouldn't matter whether the kernel recognizes sda and sdb as sdb and sda respectively should you plug them in differently. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktikynxk1ht0m-1azgipknzqvoyi-od2sxixth...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
But how can this be correct when each raid partion is linked to the HDD/Partions # cat /proc/mdstat Personalities : [raid1] [raid6] [raid5] [raid4] md3 : active raid1 sda2[0] sdb2[1] 716796096 blocks [2/2] [UU] md2 : active raid1 sda5[0] sdb5[1] 51199040 blocks [2/2] [UU] md0 : active raid1 sda1[0] sdb1[1] 513984 blocks [2/2] [UU] md1 : active raid1 sda3[0] sdb3[1] 102398208 blocks [2/2] [UU] mdadm assembles an array according to data in the superblock so it shouldn't matter whether the kernel recognizes sda and sdb as sdb and sda respectively should you plug them in differently. A good point. Noted -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c1b44c2.8010...@ionic.co.uk
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Fri, Jun 18, 2010 at 6:02 PM, Tom H tomh0...@gmail.com wrote: [snip] mdadm assembles an array according to data in the superblock so it shouldn't matter whether the kernel recognizes sda and sdb as sdb and sda respectively should you plug them in differently. so they's recognized with data in the superblock, that even uuid doesn't matter? Tao -- http://huangtao.me/ http://www.google.com/profiles/UniIsland -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimfqhxqdjrq5ljcwwfwtltwsb6ak_ydd2bzf...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Thu, Jun 17, 2010 at 09:08:41PM +0800, Huang, Tao wrote: On Thu, Jun 17, 2010 at 4:17 PM, Michal mic...@ionic.co.uk wrote: This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) I think the allocating of /dev/sdX depends on the order you plug the drives into the machine. so it changes over reconfiguring of the hardwares, which makes your labels useless. can someone confirm this? I had something like this happen on a Lenny amd64 system. The drive identifications (/dev/sdX) switched after I performed a kernel upgrade. If I booted the old kernel, they were back to normal. That's when I learned about UUID's... -Rob -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100618192626.ga10...@aurora.owens.net
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, June 16, 2010 17:30, Michal wrote: Sorry I really didnt explain my self propely; Yes I mean /dev/sde and by lable I mean get a lable machine (or somehting similar) to put a physical lable on the drive, like a sticker with text saying /dev/sde I did this in one machine and simply built my RAID1 array across two drives, disconnected a drive, booted back up check mdstat to see which one was now disconnected and labled that one, then labled the second one. It's not a brilliant way I will admit but it works perfectly well. I tested it 3 times (connecting the drive back, rebuild array, disconnecting the other drive etc) to really make sure I had labled them correctly. Ah, now I get it, I had no idea how to know which drive to put the right label on. Thanks. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/10288.91.183.48.98.1276758593.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, June 16, 2010 17:30, Michal wrote: Sorry I really didnt explain my self propely; Yes I mean /dev/sde and by lable I mean get a lable machine (or somehting similar) to put a physical lable on the drive, like a sticker with text saying /dev/sde I did this in one machine and simply built my RAID1 array across two drives, disconnected a drive, booted back up check mdstat to see which one was now disconnected and labled that one, then labled the second one. It's not a brilliant way I will admit but it works perfectly well. I tested it 3 times (connecting the drive back, rebuild array, disconnecting the other drive etc) to really make sure I had labled them correctly. Ah, now I get it, I had no idea how to know which drive to put the right label on. Thanks. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/10287.91.183.48.98.1276758573.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On 16/06/2010 19:00, Håkon Alstadheim wrote: Steven skrev: How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server. Just do ls -l /dev/disk/by-id/. The disks will have factory labels with serial-numbers to match. This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c19da2d.1090...@ionic.co.uk
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Thu, June 17, 2010 10:17, Michal wrote: On 16/06/2010 19:00, Håkon Alstadheim wrote: Just do ls -l /dev/disk/by-id/. The disks will have factory labels with serial-numbers to match. This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) Excellent, thank you both, this seems like the fastest/best way. Backups and RAID is one thing, but they're both useless if you can't recover :) Kind regards, Steven PS. I hope I fixed the duplicate mail issue now. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1366.194.7.9.50.1276768902.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
also sprach Michal mic...@ionic.co.uk [2010.06.17.1017 +0200]: This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) Instead, I suggest you stop using /dev/sdX everywhere and only use /dev/disk/by-id/*. And/or file a bug against the kernel to request that /proc/mdstat should list the ID. -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems logik ist analsadismus: gedanken werden gewaltsam durch einen engen gang gepreßt. -- frei nach lacan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100617115801.gd25...@fishbowl.rw.madduck.net
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Thu, Jun 17, 2010 at 4:17 PM, Michal mic...@ionic.co.uk wrote: This is a better way then disconnecting the drive and checking which drive was disconnected like I did, but I would still put a very easy to read label on the drive to say /dev/sdX. It would be far easier then checking a long serial number, especially if it's hard to read and you'd need to take each HDD out to check :) I think the allocating of /dev/sdX depends on the order you plug the drives into the machine. so it changes over reconfiguring of the hardwares, which makes your labels useless. can someone confirm this? Tao -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimc8j0kf165dwkqryugtckavjopycu3agoea...@mail.gmail.com
Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
Hope some one finds this helpful :-) --Siju Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault. = ** Actual screen shot from terminal of steps taken during rebuild on 10-June-2010 on Debian Lenny ( Linux )** 1) Check the partitions layout on the current hard disk srv1:~# fdisk /dev/sda The number of cylinders for this disk is set to 60801. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) software that runs at boot time (e.g., old versions of LILO) 2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Command (m for help): p Disk /dev/sda: 500.1 GB, 500107862016 bytes 255 heads, 63 sectors/track, 60801 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0xdd6e Device Boot Start End Blocks Id System /dev/sda1 1 122 979933+ fd Linux raid autodetect /dev/sda2 1231338 9767520 fd Linux raid autodetect /dev/sda313392554 9767520 fd Linux raid autodetect /dev/sda42555 60801 467869027+ fd Linux raid autodetect Command (m for help): quit srv1:~# 2) Create identical partitions on the new disk using 'fdisk'. Partition Id should be 'fd' for all RAID partitions. The resulting layout should look like. srv1:~# fdisk /dev/sdb The number of cylinders for this disk is set to 60801. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) software that runs at boot time (e.g., old versions of LILO) 2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Command (m for help): p Disk /dev/sdb: 500.1 GB, 500107862016 bytes 255 heads, 63 sectors/track, 60801 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0xe3a3a447 Device Boot Start End Blocks Id System /dev/sdb1 1 122 979933+ fd Linux raid autodetect /dev/sdb2 1231338 9767520 fd Linux raid autodetect /dev/sdb313392554 9767520 fd Linux raid autodetect /dev/sdb42555 60801 467869027+ fd Linux raid autodetect Command (m for help): q srv1:~# 3) Check the current RAID status srv1:~# cat /proc/mdstat Personalities : [raid1] md3 : active raid1 sda4[1] 467868928 blocks [2/1] [_U] md2 : active raid1 sda3[1] 9767424 blocks [2/1] [_U] md1 : active raid1 sda2[1] 9767424 blocks [2/1] [_U] md0 : active raid1 sda1[1] 979840 blocks [2/1] [_U] unused devices: srv1:~# 4) Rebuild the arrays and check thr status srv1:~# mdadm -a /dev/md0 /dev/sdb1 mdadm: added /dev/sdb1 srv1:~# mdadm -a /dev/md1 /dev/sdb2 mdadm: added /dev/sdb2 srv1:~# mdadm -a /dev/md2 /dev/sdb3 mdadm: added /dev/sdb3 srv1:~# mdadm -a /dev/md3 /dev/sdb4 mdadm: added /dev/sdb4 srv1:~# cat /proc/mdstat Personalities : [raid1] md3 : active raid1 sdb4[2] sda4[1] 467868928 blocks [2/1] [_U] [] recovery = 0.0% (285440/467868928) finish=54.5min speed=142720K/sec md2 : active raid1 sdb3[0] sda3[1] 9767424 blocks [2/2] [UU] md1 : active raid1 sdb2[0] sda2[1] 9767424 blocks [2/2] [UU] md0 : active raid1 sdb1[0] sda1[1] 979840 blocks [2/2] [UU] unused devices: srv1:~# 5) Install grub on the MBR of new hard disk srv1:~# grub-install /dev/sdb Searching for GRUB installation directory ... found: /boot/grub Installation finished. No error reported. This is the contents of the device map /boot/grub/device.map. Check if this is correct or not. If any of the lines is incorrect, fix it and re-run the script `grub-install'. (hd0) /dev/sda (hd1) /dev/sdb srv1:~# -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktinfezni1ezq0vkdo4ryhh7uolrugaaje9_l1...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
also sprach Siju George sgeorge...@gmail.com [2010.06.16.1313 +0200]: 2) Create identical partitions on the new disk using 'fdisk'. sfdisk -d /dev/sda | sfdisk /dev/sdb -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems i always choose my friends for their good looks and my enemies for their good intellects. man cannot be too careful in his choice of enemies. -- oscar wilde -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100616111821.gb5...@piper.oerlikon.madduck.net
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, Jun 16, 2010 at 4:48 PM, martin f krafft madd...@debian.org wrote: also sprach Siju George sgeorge...@gmail.com [2010.06.16.1313 +0200]: 2) Create identical partitions on the new disk using 'fdisk'. sfdisk -d /dev/sda | sfdisk /dev/sdb oh thanks :-) I did it manually using fdisk --Siju -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktilgnr-wlrnoqkmxlpyou0ep_vb-h3vlm4o6g...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
also sprach Siju George sgeorge...@gmail.com [2010.06.16.1322 +0200]: sfdisk -d /dev/sda | sfdisk /dev/sdb oh thanks :-) I did it manually using fdisk Manually is for Mac users. ;) -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems work like you don't need the money love like you have never been hurt dance like there's nobody watching digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/)
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, Jun 16, 2010 at 5:06 PM, martin f krafft madd...@debian.org wrote: also sprach Siju George sgeorge...@gmail.com [2010.06.16.1322 +0200]: sfdisk -d /dev/sda | sfdisk /dev/sdb oh thanks :-) I did it manually using fdisk Manually is for Mac users. ;) these days every one has left windows and are picking on Mac ? :-) -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimgxdr683grsg4piso_pivu6pjwjkaspjgsm...@mail.gmail.com
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
also sprach Siju George sgeorge...@gmail.com [2010.06.16.1402 +0200]: Manually is for Mac users. ;) these days every one has left windows and are picking on Mac ? :-) Reinstalling is for Windows users. -- .''`. martin f. krafft madd...@d.o Related projects: : :' : proud Debian developer http://debiansystem.info `. `'` http://people.debian.org/~madduckhttp://vcs-pkg.org `- Debian - when you have better things to do than fixing systems the reason the mainstream is thought of as a stream is because it is so shallow. digital_signature_gpg.asc Description: Digital signature (see http://martin-krafft.net/gpg/)
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, June 16, 2010 13:13, Siju George wrote: Hope some one finds this helpful :-) --Siju Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault. = Thanks, this might prove useful. However I do have a question... which might be just as important. How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/23452.91.183.48.98.1276695172.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
Thanks, this might prove useful. However I do have a question... which might be just as important. How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server. One way is to label the disks themselves so you simply do; cat /proc/mdstat which might say /dev/sd3 is down. Open the case, look for the disk labled /dev/sde and replace it. If you have LED's like servers have (probably not) they can be a fiddle to get working but it's possible -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c18d5fb.2000...@ionic.co.uk
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, June 16, 2010 13:13, Siju George wrote: Hope some one finds this helpful :-) --Siju Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault. = Thanks, this might prove useful. However I do have a question... which might be just as important. How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/23451.91.183.48.98.1276695152.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On Wed, June 16, 2010 15:47, Michal wrote: One way is to label the disks themselves so you simply do; cat /proc/mdstat which might say /dev/sd3 is down. Open the case, look for the disk labled /dev/sde and replace it. If you have LED's like servers have (probably not) they can be a fiddle to get working but it's possible No LED's for drives, it already has them for every pci slot, looks like a Christmas tree :) I think you meant /dev/sde instead of sd3, right? If not, please correct me. If I'm not mistaken, mdadm will report the broken drive, then I have to look for the drive that corresponds to the 4th sata slot on the motherboard. That's part of my issue, can I be sure that the drive connected to port 4 is /dev/sde? It's not a problem for the other 2 drives, as they differ in capacity, but these 4 are exactly the same size. Also how accurate is mdadm in identifying the failed drive? As there are only 2 in an array, there is only 1 copy of the data to compare to. It also seems my last message was sent twice, sorry about that. -- Rarely do people communicate; they just take turns talking. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/23813.91.183.48.98.1276699820.squir...@stevenleeuw.kwik.to
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
On 16/06/2010 15:50, Steven wrote: On Wed, June 16, 2010 15:47, Michal wrote: One way is to label the disks themselves so you simply do; cat /proc/mdstat which might say /dev/sd3 is down. Open the case, look for the disk labled /dev/sde and replace it. If you have LED's like servers have (probably not) they can be a fiddle to get working but it's possible No LED's for drives, it already has them for every pci slot, looks like a Christmas tree :) I think you meant /dev/sde instead of sd3, right? If not, please correct me. If I'm not mistaken, mdadm will report the broken drive, then I have to look for the drive that corresponds to the 4th sata slot on the motherboard. That's part of my issue, can I be sure that the drive connected to port 4 is /dev/sde? It's not a problem for the other 2 drives, as they differ in capacity, but these 4 are exactly the same size. Also how accurate is mdadm in identifying the failed drive? As there are only 2 in an array, there is only 1 copy of the data to compare to. It also seems my last message was sent twice, sorry about that. Sorry I really didnt explain my self propely; Yes I mean /dev/sde and by lable I mean get a lable machine (or somehting similar) to put a physical lable on the drive, like a sticker with text saying /dev/sde I did this in one machine and simply built my RAID1 array across two drives, disconnected a drive, booted back up check mdstat to see which one was now disconnected and labled that one, then labled the second one. It's not a brilliant way I will admit but it works perfectly well. I tested it 3 times (connecting the drive back, rebuild array, disconnecting the other drive etc) to really make sure I had labled them correctly. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c18ee14.6060...@ionic.co.uk
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
Use smartctl from the smartmontools package. If mdadm says that /dev/sdc (or cat /proc/mdstat) is at fault then use smartctl -a /dev/sdc and it will print out all kinds of info on the drive including its serial number which should be on a sticker on the case of the drive. The programs included with smartmontools might have warned you of an impending failure. I have a smart self long test run om my drives 2 times a week. *...Bob* On 06/16/2010 09:32 AM, Steven wrote: On Wed, June 16, 2010 13:13, Siju George wrote: Hope some one finds this helpful :-) --Siju Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault. = Thanks, this might prove useful. However I do have a question... which might be just as important. How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server.
Re: Rebuilding RAID 1 Array in Linux with a new hard disk after a disk fault - Howto with screen shots
Steven skrev: How to identify which drive has failed in an array? I have 6 disks, 4 are used in raid (mdadm), the other 2 contain /boot, / and /home. /dev/sdc /dev/sdd /dev/sde /dev/sdf Each have 1 partition. /dev/md0 (raid 1) consists of /dev/sdc1 and /dev/sdd1 /dev/md1 (raid 1) consists of /dev/sde1 and /dev/sdf1 If a drive fails, how do I know which drive? This is a desktop system, not a server. Just do ls -l /dev/disk/by-id/. The disks will have factory labels with serial-numbers to match. -- Håkon Alstadheim / N-7510 Skatval / email:ha...@alstadheim.priv.no tlf: 74 82 60 27 mob: 47 35 39 38 http://alstadheim.priv.no/hakon/ spamtrap: finnesi...@alstadheim.priv.no -- 1 hit you are out -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c191159.8080...@alstadheim.priv.no
