RAID
openmediavault uses linux software RAID driver (MD) and the mdadm utility to create arrays [1]. Arrays created in any other linux distro should be recognized immediately by the server. In most cases you can skip to the filesystem array and proceed to mount to integrate the filesystem into the database.
Overview
The grid panel shows all currently available MD arrays. There is no internal database section for RAID arrays, so every array that is assembled in the server should be displayed here.
- Create
The following table displays levels available in the web interface:
RAID Level
Name
Min. Disks
Redundancy
Growable
Linear
JBOD
2
No
No
0
Stripe
2
No
No
1
Mirror
2
Yes
Yes
5
RAID5
3
Yes
Yes
6
RAID6
4
Yes
Yes
10
Stripped Mirror
4
Yes
No
Note
RAID4 and FAULTY levels are not supported in the web interface.
RAID1+0 is possible by stripping two mirrors. The create window should display both mirrors if they do not have any filesystem signatures yet.
- Detail
Displays extended information of the array, the output comes from mdadm –detail /dev/mdX
- Grow
Add disk(s) into the array.
- Recover
If the array comes from another linux server you can use this button to reassemble the array in the current server
- Remove
This is used to remove failed disks, in case one needs be replaced.
- Delete
Stop the array and zero the superblock of all devices conforming the array (script /usr/sbin/omv-rmraid). Use with caution.
Mdadm works better with unpartitioned disks, plain raw block devices. Before creating MD RAID in your system make sure disks are clean before. In the physical disk section you can perform a quick or full wipe. Quick wipe is enough to delete partition tables.
Degraded array creation in not possible in the web interface, however the array can be created in terminal using mdadm if you want for example to convert a RAID from level 1 to 5 or 6.
Mail notifications are integrated for mdadm, these are sent everytime an array enters degraded state.
Growing
Before growing array is better to clean the partition table of the new disk, specially if the disk was used before in another mdadm array, erase also the superblock:
$ mdadm --zero-superblock /dev/sdX
After adding a disk to the array, the re-synching process will begin immediately. Depending on the size of the disks this process can take several hours or even days, this is because mdadm tries to balance resources and keep usability of the system not using high CPU and RAM. To speed the process:
$ echo ${value} > /proc/sys/dev/raid/speed_limit_min #value is interpreted as kbytes/seconds
After synching is finish is necessary to expand the filesystem using the resize button.
Warning
If your MD array and filesystem was created with openmediavault or Debian before December 2014, then please read here.
Do not use RAID arrays in production with drives connected via USB, neither hubs or different ports. This includes low power devices that do not have a SATA controller, e.g. Raspberry Pi, Pogoplugs and any low entry ARM SBC.