Mdadm
There are a few options for software RAID on Gnu Linux. Among them is BtrFS and ZFS (the latter of which has licensing issues, so use discretion before using)., however today I will focus on mdadm. This is historically the oldest software raid, therefore should be better vetted, although its performance may be less of that of the first two mentioned - for simple servers, mdadm might be the most stable choice. However, after testing it in practice I found that it was not stable on 1 out of 2 machines. The general rule that 'software RAID is not trustworthy' might apply here. Generally, go with hardware RAID.
Creation of RAID
Will not be covered here (yet). You must create the partition tables. Create the raid with mdadm. mkfs.ext4 on the raid partition. Add mdadm to grub config. Reinstall grub. Details may be provided later.
Usage
Status / Details
# sudo mdadm -D /dev/md127
/dev/md127:
Version : 1.2
Creation Time : Fri Feb 1 01:00:25 2019
Raid Level : raid1
Array Size : 57638912 (54.97 GiB 59.02 GB)
Used Dev Size : 57638912 (54.97 GiB 59.02 GB)
Raid Devices : 3
Total Devices : 2
Persistence : Superblock is persistent
Update Time : Fri Feb 1 02:40:44 2019
State : clean, degraded
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0
Name : devuan:root
UUID : 83a8dc03:802a4129:26322116:c2cfe1d4
Events : 82
Number Major Minor RaidDevice State
- 0 0 0 removed
1 8 17 1 active sync /dev/sdb1
2 8 33 2 active sync /dev/sdc1
(here you can see, one was removed. It auto removes, when unplugged)
Add Drive to Array
This assumes you have already created the RAID. This is for adding to an existing array.
# mdadm --add /dev/md127 /dev/sda1
NOTE: If you setup 2 hdds, in a raid, and want to add a third, if you just --add, it will show up as a spare...
if you do mdadm --grow /dev/md127 -raid-devices=3 then the third might be active sync (what we want) note that the --grow, seems to allow for parameter changes after you have already created the raid. you can also specify the exact same command, raid-devices=3 in the setup of the raid (see install doc). Note that if you lose a drive, you can simply add it.
NOTE: don't worry about mkfs.ext4 on the raid members, after initial setup. The RAID will manage that.
NOTE: if you have a new drive and need to copy the hdd partition tables: https://unix.stackexchange.com/questions/12986/how-to-copy-the-partition-layout-of-a-whole-disk-using-standard-tools or aka
(FOR MBR ONLY) Save: sfdisk -d /dev/sda > part_table Restore: sfdisk /dev/NEWHDD < part_table (FOR GPT:) # Save MBR disks sgdisk --backup=/partitions-backup-$(basename $source).sgdisk $source sgdisk --backup=/partitions-backup-$(basename $dest).sgdisk $dest # Copy $source layout to $dest and regenerate GUIDs sgdisk --replicate=$dest $source sgdisk -G $dest
Add HDD back to RAID (to fix failed drive)
# mdadm --add /dev/md127 /dev/sda1
mdadm: added /dev/sda1
# sudo mdadm -D /dev/md127
/dev/md127:
Version : 1.2
Creation Time : Fri Feb 1 01:00:25 2019
Raid Level : raid1
Array Size : 57638912 (54.97 GiB 59.02 GB)
Used Dev Size : 57638912 (54.97 GiB 59.02 GB)
Raid Devices : 3
Total Devices : 3
Persistence : Superblock is persistent
Update Time : Fri Feb 1 02:41:43 2019
State : clean, degraded, recovering
Active Devices : 2
Working Devices : 3
Failed Devices : 0
Spare Devices : 1
Rebuild Status : 0% complete
Name : devuan:root
UUID : 83a8dc03:802a4129:26322116:c2cfe1d4
Events : 92
Number Major Minor RaidDevice State
3 8 1 0 spare rebuilding /dev/sda1
1 8 17 1 active sync /dev/sdb1
2 8 33 2 active sync /dev/sdc1
You can check on the rebuild partway:
Rebuild Status : 6% complete
Name : devuan:root
UUID : 83a8dc03:802a4129:26322116:c2cfe1d4
Events : 103
Number Major Minor RaidDevice State
3 8 1 0 spare rebuilding /dev/sda1
1 8 17 1 active sync /dev/sdb1
2 8 33 2 active sync /dev/sdc1
You can also watch it via /proc/
watch -n1 cat /proc/mdstat
Every 1.0s: cat /proc/mdstat
server: Fri Feb 1 02:43:24 2019
Personalities : [raid1] [linear] [multipath] [raid0] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sda1[3] sdb1[1] sdc1[2]
57638912 blocks super 1.2 [3/2] [_UU]
[==>..................] recovery = 11.2% (6471936/57638912) finish=13.2min speed=64324K/sec
unused devices: <none>
WARNING: Reinstall grub on the new drive again as well afterwards.
Email Notifications
Test emails on mdadm.. first configure email however you prefer (i currently use ssmtp).
edit /etc/mdadm/mdadm.conf; put the email recipient in mailadd variable
then
# mdadm --monitor --scan --test --oneshot
should send an email
(Be careful that updates don't break this file. I see it's generated by mkconf.)
https://ubuntuforums.org/showthread.php?t=1185134
for more details on email sending
External Links
The section about degraded disks: https://help.ubuntu.com/lts/serverguide/advanced-installation.html.en
General partition tips: https://github.com/zfsonlinux/zfs/wiki/Debian-Stretch-Root-on-ZFS
Configuring Email w/ssmtp: https://wiki.zoneminder.com/SMS_Notifications
Postscript
I had one of two the mdadm raids give me errors, and I gave up on it. RAID is difficult business to deal with, and maybe ZFS is better than mdadm. But I prefer to not use RAID when possible. Instead, I like to have drop in replacement machines (like datto's idea), which is cheap and easy, given computer hardware is now commodity. There are a few problems with RAID (both hardware and software) such as:
- Hardware RAID writes special snowflake data to HDDs that can't easily be read by the OS.
- Hardware RAID is indecipherable to end users (requires a technician to fix)
- Software RAID is notorious for being unstable
- All the RAID in the world won't help, if there is a hardware failure besides the HDDs (i.e. motherboard fails).
- Probably there is a reliable Hardware RAID setup, but I'm not enough 'in the know' to know what it is.
- Given the R / S pun, with the word RAID you have AIDS. The hint may be in the language.
- Hard drives are mostly reliable (and must be better than they were in the 2000's) (Spinning drives, not SSDs).
| ||||||||||||||||||
