I would love to know what you use your array for (i.e. pictures, videos, documents, backups, VM storage, etc.) This would make recommending a RAID setup much easier.
Printable View
I would need to know what you intent to use the array for, but off the top of my head, I would suggest going with RAID 6 because you have so many drives - but that is only from a data storage standpoint.
Agreed. I use my array (RAID5) for backups, file storage, VM storage.. and all that, but RAID isn't a "one size fits all" type of thing.
Thanks to you my situation seems to improve.
Current situation: All drives are up and running
Raid assembling worked on the second try. Had to stop and assemble two of the arrays to get them up.
mdstat ist looking like this
I keep my fingers crossed that this is going to work.Code:root@mediacenter:/home/ludwig# more /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md1 : active raid5 sdd1[1] sdb1[0] sdc1[3]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [_UU]
resync=DELAYED
md0 : active raid1 sdd8[1] sdc8[3] sdb8[0]
116390840 blocks super 1.2 [3/2] [U_U]
resync=DELAYED
md7 : active raid5 sdd7[1] sdc7[3] sdb7[4]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [U_U]
[>....................] recovery = 2.1% (5739036/262142784) finish=103.3min speed=41349K/sec
md6 : active raid5 sdd6[1] sdc6[3] sdb6[0]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [U_U]
resync=DELAYED
md5 : active raid5 sdd5[1] sdc5[3] sdb5[0]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [_UU]
resync=DELAYED
md4 : active raid5 sdd4[1] sdb4[0] sdc4[3]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [_UU]
resync=DELAYED
md3 : active raid5 sdd3[1] sdc3[3] sdb3[4]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [U_U]
resync=DELAYED
md2 : active raid5 sdd2[4] sdc2[3] sdb2[0]
524285568 blocks super 1.2 level 5, 64k chunk, algorithm 2 [3/2] [U_U]
resync=DELAYED
unused devices: <none>
And thanks again for your help
The main usage for the system is a file server for media files like music, video, etc. It also has a TV card for recording purposes. So the majority of the files are quite large.
There's also the mirror set that contains very important document files shared in the internal network and automatically backed up in the cloud.
There are also some other services like a web server or subversion but they are used not much compared to samba.
Since it is likely to happen to run out of space one goal would be easy extensability. But currently I have only 4 SATA ports.
Does this provide a picture sufficient for a proposal?
Thanks
dolittle
I love mdadm and use it everyday, but for home media storage, I've personally switched over to Snapraid. Each disk is independent and is mounted like a normal ext4 filesystem. This means that even if you lose more parity than your array supports, you still only lose the data on that one disk.
I have a tutorials for both mdadm or snapraid in my signature if you are interested. Either way, Snapraid or mdadm, I'd use RAID6, and use one big array, no need for a bunch of little arrays, or even LVM. You can expand either one of these solutions without the need for LVM.
If you have an extra PCI-Express x8 or x16 slot, you can easily add a card like the IBM m1015 or BR10i flashed to IT mode and add up to 8 SATA ports (you'll need forward breakout cables for either of these to hook up to disks).
I haven't used Snapraid at all, but I do +1 using RAID6 with that number of disks.
I've been running RAID5 with 3 drives, but anything with 5 or more would be better off running RAID6.