RAID

From Indie IT Wiki
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Software

Introduction

https://wiki.archlinux.org/index.php/RAID

Cancel A RAID Resync

sudo /usr/share/mdadm/checkarray -x --all

HOWTO: Rename A RAID Array

http://askubuntu.com/questions/63980/how-do-i-rename-an-mdadm-raid-array

MD Device Number Suddenly Changes

FIX - http://ubuntuforums.org/showthread.php?t=1764861

BEST PRACTICE - http://askubuntu.com/questions/211180/ubuntu-server-12-04-mdadm-device-number-suddenly-changes

Change Preferred Minor MD Superblock On Software RAID Hard Disk

This only works on 0.9 Software Raid metadata.

A hard disk drive failure on one of my servers prompted a manual software RAID 1 check and resync. Using good old SystemRescueCD, this was done but in doing so it inadvertently changed the preferred minor md superblock number on the array's hard disk drives from 3 to 124.

mdadm --misc --examine /dev/sda3 |grep 'Preferred Minor'
Preferred Minor : 124

To fix this, stop the array and reassemble it again but update the superblock as it runs. Make sure you declare which hard disk drive partitions you are using as you assemble:-

mdadm --verbose --misc --stop /dev/md124
mdadm --verbose --assemble --update=super-minor --run /dev/md3 /dev/sda3 /dev/sdb3

To check it has worked, run the following 2 commands:-

mdadm --misc --examine /dev/sda3 |grep 'Preferred Minor'
Preferred Minor : 3
cat /proc/mdstat
md3 : active raid1 sdb3[0] sda3[1]
     10490368 blocks [2/2] [UU]

HOWTO: Install MDADM Without Postfix

sudo apt-get install mdadm --no-install-recommends

HOWTO: Send Test Email

mdadm --monitor --scan --test --oneshot

ERROR: ubuntu boot failed device or resource busy

Possibly a hard disk fault or the fact that the initraamfs has not given enough time for the RAID devices to assemble properly.

The fix is to add a delay to the boot process.

  1. Hold the right SHIFT key down to bring up the GRUB menu
  2. Select 'Advanced options for Ubuntu...'
  3. Choose 'Ubuntu Recovery Mode'
  4. Press the e key
  5. Add rootdelay=90 on the line just before the root= part
  6. Press F10 to boot into Recovery Mode
  7. FSCK the disks
  8. Drop to root prompt
  9. reboot
  10. Repeat for normal 'Ubuntu' line
  11. When properly booted, do the following...
sudo -i
echo "sleep 60" > /etc/initramfs-tools/scripts/init-premount/delay_for_raid_array_to_build_before_mounting
chmod a+x /etc/initramfs-tools/scripts/init-premount/delay_for_raid_array_to_build_before_mounting
update-initramfs -u
reboot

Thanks - http://ubuntuforums.org/showthread.php?t=2241430

Thanks - http://www.linuxtopia.org/online_books/linux_kernel/kernel_configuration/re58.html

HOWTO: Set Up A 3TB Disk

With hard disk drives larger than 2TB, you need to use different software and commands to set up Linux Software RAID 1.

The crucial fact is that they use the new GPT or (GUID Partition Table) layout.

You use parted instead of fdisk, and sgdisk instead of sfdisk.

This will make a disk with 3 partitions...

  1. bios_grub (boot) ~3Mb
  2. raid (swap) ~1Gb
  3. raid (rootfs) ~3Tb

Stage 1 - Partition The Drives (parted)

parted -a optimal /dev/sda
(parted) mklabel gpt
(parted) unit mib
(parted) mkpart primary 1 3
(parted) name 1 grub
(parted) set 1 bios_grub on
(parted) mkpart primary 3 1000
(parted) name 2 swap
(parted) set 2 raid on
(parted) mkpart primary 1000 -1
(parted) name 3 rootfs
(parted) set 3 raid on
(parted) align-check optimal 1
(parted) align-check optimal 2
(parted) align-check optimal 3
(parted) print
Number  Start   End     Size    File system  Name    Flags
1       1049kB  3146kB  2097kB               grub    bios_grub
2       3146kB  1049MB  1045MB               swap    raid
3       1049MB  3001GB  3000GB               rootfs  raid
(parted) quit

Stage 2 - Copy Partitions To Other Drives (sgdisk)

sgdisk --backup=table /dev/sda
sgdisk --load-backup=table /dev/sdb
sgdisk -G /dev/sdb
sgdisk --backup=table /dev/sda
sgdisk --load-backup=table /dev/sdc
sgdisk -G /dev/sdc

Stage 3 - Check Partitions

lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda      8:0    0   2.7T  0 disk
├─sda1   8:1    0     2M  0 part
├─sda2   8:2    0   997M  0 part
└─sda3   8:3    0   2.7T  0 part
sdb      8:16   0   2.7T  0 disk
├─sdb1   8:17   0     2M  0 part
├─sdb2   8:18   0   997M  0 part
└─sdb3   8:19   0   2.7T  0 part
sdc      8:32   0   2.7T  0 disk
├─sdc1   8:33   0     2M  0 part
├─sdc2   8:34   0   997M  0 part
└─sdc3   8:35   0   2.7T  0 part

Stage 4 - Set Up RAID Arrays (mdadm)

mdadm --create --verbose /dev/md0 --level=mirror --raid-devices=2 /dev/sda2 /dev/sdb2 --spare-devices=1 /dev/sdc2
mdadm --create --verbose /dev/md1 --level=mirror --raid-devices=2 /dev/sda3 /dev/sdb3 --spare-devices=1 /dev/sdc3

Stage 5 - Format (Ubuntu Installer or mkfs.ext4)


...

Deleting The RAID Superblock / Wiping A Hard Disk Drive

  • Remove the superblock
mdadm --zero-superblock /dev/sdX1

If you get this error...

mdadm: Couldn't open /dev/sdX1 for write - not zeroing

Then stop the arrays which have already been started by mdadm...

mdadm --manage --stop /dev/md127
mdadm --manage --stop /dev/md126
mdadm --manage --stop /dev/md125
mdadm --manage --stop /dev/md124

Then stop the mdadm service...

/etc/init.d/mdadm stop

And try again...

mdadm --zero-superblock /dev/sdX1
  • Delete the partitions
dd if=/dev/zero of=/dev/sdX bs=512 count=1
  • Securely wipe the data
time shred -n 1 -vz /dev/sdX

Tips To Speed Up RAID Rebuilding And Resync

http://www.cyberciti.biz/tips/linux-raid-increase-resync-rebuild-speed.html

Remove mdadm Rebuild Speed Restriction

sudo cp /etc/sysctl.conf /etc/sysctl.conf_ORIG
sudo nano /etc/sysctl.conf
     dev.raid.speed_limit_max = 51200

Where 51200 is the value to the speed in KB/s you would like to use, in this case 50 MB/s.

Adding Bitmap Indexes To mdadm

Including a bitmap index to a mdadm before rebuilding the array can also speed up the process of rebuilding.


sudo mdadm --grow --bitmap=internal /dev/md0

After the array has been rebuilt the bitmap index can be removed:


mdadm --grow --bitmap=none /dev/md0

NOTE: The above example assumes the array can be found at md0.

Thanks go to James Coyle's article here.

Replacing A Failed Hard Drive In A Software RAID1 Array

Oh no, the second hard drive has broken and we need to swap it!

  • Mark the hard drive as failed in all arrays
mdadm --manage /dev/md0 --fail /dev/sdb1
mdadm --manage /dev/md1 --fail /dev/sdb2
  • Remove the hard drive from all arrays
mdadm --manage /dev/md0 --remove /dev/sdb1
mdadm --manage /dev/md1 --remove /dev/sdb2
  • Power down the system
poweroff
  • Replace the hard drive and boot the system
  • Copy the partitions from the current disk to the new disk
sfdisk -d /dev/sda | sfdisk /dev/sdb
  • Check the partitions match
fdisk -l /dev/sda /dev/sdb
  • Add the new hard drive to all arrays
mdadm --manage /dev/md0 --add /dev/sdb1
mdadm --manage /dev/md1 --add /dev/sdb2
  • Check when finished
cat /proc/mdstat
  • Reboot just for good measure
reboot

http://www.howtoforge.com/replacing_hard_disks_in_a_raid1_array

Scan, Unmount, Stop and Start A RAID Array

Scan

mdadm --examine --scan

Unmount

umount /dev/md0

Stop

mdadm --misc --verbose --stop /dev/md0

Start

mdadm --assemble --verbose --run /dev/md0

Scan And Start Array

sudo mdadm --examine --scan                    
ARRAY /dev/md/0  metadata=1.2 UUID=a3ae9396:8085fdd8:ce69dd19:28a50bb8 name=sysresccd:0 spares=1
ARRAY /dev/md/1  metadata=1.2 UUID=3ce66ea0:21de7908:856e1488:18471642 name=sysresccd:1 spares=1

sudo mdadm --assemble --scan
mdadm: /dev/md/1 has been started with 2 drives and 1 spare.
mdadm: /dev/md/0 has been started with 2 drives and 1 spare.

sudo cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sda2[0] sdc2[2](S) sdb2[1]
      1020352 blocks super 1.2 [2/2] [UU]
md1 : active raid1 sda3[0] sdc3[2](S) sdb3[1]
      2929110464 blocks super 1.2 [2/2] [UU]
      bitmap: 0/22 pages [0KB], 65536KB chunk
unused devices: <none>

Use Setup RAID Array For Swap And Root Filesystem

sudo mkswap --label swap /dev/md0
Setting up swapspace version 1, size = 996.4 MiB (1044836352 bytes)
LABEL=swap, UUID=85ec8988-da79-4747-9025-32a185b1cb95

swapon -L swap

sudo free -m
              total        used        free      shared  buff/cache   available
Mem:           3860          43        3671           2         145        3773
Swap:           996           0         996

sudo mkfs.ext4 -L rootfs /dev/md1
mke2fs 1.42.13 (17-May-2015)
Creating filesystem with 732277616 4k blocks and 183074816 inodes
Filesystem UUID: 38f6c857-7c5b-45e2-92ef-a4da101b9b9e
Superblock backups stored on blocks: 
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, 
102400000, 214990848, 512000000, 550731776, 644972544
Allocating group tables: done                            
Writing inode tables: done                            
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

Start Array With 1 Drive To Access Data

Find names of arrays to use...

sudo mdadm --examine --scan

Assemble array with just one drive...

sudo mdadm --assemble --run /dev/md1 /dev/sdc1
sudo mdadm --assemble --run /dev/md2 /dev/sdc2
sudo mdadm --assemble --run /dev/md3 /dev/sdc3
sudo mdadm --assemble --run /dev/md4 /dev/sdc4

Check filesystem type...

sudo blkid /dev/md1
sudo blkid /dev/md2
sudo blkid /dev/md3
sudo blkid /dev/md4

Create mount directories...

sudo mkdir /mnt/md{1,2,3,4}

Mount as required...

sudo mount -v -t ext3 /dev/md3 /mnt/md3

Read as required...

sudo ls -lah /mnt/md3

When you have finished...

sudo sync
sudo umount /mnt/md3
sudo mdadm --stop --scan

...job, done.

Extending a RAID Device

To add a new device to an existing array, use the command in the following form as root:

mdadm raid_device --add component_device

This will add the device as a spare device.

To grow the array to use this device actively, type the following at a shell prompt:

mdadm --grow raid_device --raid-devices=number

Assume the system has an active RAID device, /dev/md3, with the following layout (that is, the RAID device created in Example 5.2, “Creating a new RAID device”):

mdadm --detail /dev/md3 | tail -n 3
   Number   Major   Minor   RaidDevice State
      0       8        1        0      active sync   /dev/sda1
      1       8       17        1      active sync   /dev/sdb1

Also assume that a new SCSI disk drive, /dev/sdc, has been added and has exactly one partition. To add it to the /dev/md3 array, type the following at a shell prompt:

mdadm /dev/md3 --add /dev/sdc1
mdadm: added /dev/sdc1

This will add /dev/sdc1 as a spare device. To change the size of the array to actually use it (all the time), type:

mdadm --grow /dev/md3 --raid-devices=3

https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/5/html/Deployment_Guide/s2-raid-manage-extending.html

Hardware

How to read a RAID1 hard disk if you can’t use the RAID Controller?

http://robert.penz.name/4/how-to-read-a-raid1-hard-disk-if-you-cant-use-the-raid-controller/

Miscellaneous

HP Smart RAID Array HP 410i

http://hwraid.le-vert.net/wiki/SmartArray

http://sysadm.pp.ua/linux/hpraid-monitoring.html

http://downloads.linux.hp.com/SDR/downloads/MCP/Ubuntu/dists/

http://downloads.linux.hp.com/SDR/downloads/MCP/Ubuntu/pool/non-free/?C=M;O=D

wget http://downloads.linux.hp.com/SDR/downloads/MCP/GPG-KEY-MCP
sudo apt-key add GPG-KEY-MCP
apt-get update
apt-get install hpacucli

A basic report showing the status of your raid arrays is:

hpacucli ctrl all show config

http://www.randomhacks.co.uk/installing-hp-proliant-support-pack-psp-on-ubuntu-12-04/

http://www.datadisk.co.uk/html_docs/redhat/hpacucli.htm

Investigation

server:~# lspci |grep -i 'raid'
02:05.0 RAID bus controller: LSI Logic / Symbios Logic MegaRAID (rev 01)
server:~# lsmod |grep -i raid |sort
async_memcpy            2304  1 raid456
async_tx                6316  3 raid456,async_xor,async_memcpy
async_xor               3520  1 raid456
md_mod                 67036  6 raid10,raid456,raid1,raid0,multipath,linear
megaraid_mbox          25872  2 
megaraid_mm             8284  1 megaraid_mbox
raid0                   6368  0 
raid10                 18560  0 
raid1                  18016  0 
raid456               117264  0 
scsi_mod              129356  7 sg,sd_mod,libata,mptspi,mptscsih,scsi_transport_spi,megaraid_mbox
xor                    14696  2 raid456,async_xor
server:~# ll /dev/megadev0 
crw-rw---- 1 root root 10, 60 2013-04-25 13:55 /dev/megadev0

http://wiki.debian.org/LinuxRaidForAdmins

http://hwraid.le-vert.net/