Easy way to add more HDD

Node Operators FAQ
21

I am aware that this is an older thread, but I would like to lean towards the software RAID1 solution (pair of HDDs with the same capacity) + LVM, with which I have been working for long time, I solved a lot of problems and did not lose any data (that is, at least I don’t know about that :slight_smile: ).

I use different pairs of disks connected in RAID1. I connected these MD disks to the necessary logical volumes via LVM and I’m just allocating the necessary space.

If I run out of free unallocated space, I simply release the PV (= RAID1 multi disk) and move the data somewhere else. After releasing, I disconnect the small disks and connect the larger ones. I will build RAID1 and return it to LVM as bigger PV.

When I start to run out of space on the logical volume for Storj, I allocate more (lvextend) …

Storj still sees one logical volume, which thanks to EXT4 and LVM I can “inflate” during full load.

Of course, this approach needs preparation (especially the boot partition) :frowning: . It can also be done “while driving”, but it is more complicated.

I read many topics about “bad RAID”, but I really not spot any problems with it in my operations time. Maybe RAID5/6 should be faulty more like pure mirror. I dont know …

I like to have more controll over my storage server, so I chose maybe complicated solution, but working perfectlly for me :).

22

The problem is appear only when you start to use RAID0 (stripe volume, JBOD, MergeFS, etc.) - with one disk failure the whole node is lost.
But in case of RAID1 (mirror), there is no problem - you protected at least from one disk failure. However, it will not help if both drive dies.

23

@Alexey True and true :slight_smile: .
Sometimes some drives get “current pending sector” and in this case I move whole PV (all data on RAID1 array) to another free space in VG. One time I had damaged 3 drives at once, but luckly, all drives was from different arrays :wink:.
I hope, that when crash one drive in pair, stay only that one drive and not crash both drives at once.
I still have luck and when one drive crashed, I am still able move data safe and repair array :grin:.

I know, that is maybe waste of space, but I like my good sleeping :wink: .

24

Then they should be from a different batches, or at least purchased in a different time.

25

Yop, in many in my arrays I have totally different drives :wink: . Different vendor, rotation speed, … but equal capacity :grin: .

E.g.
md6 (1.5TB) =
Western Digital Green (WDC WD15EARX-00PASB0)
+
SAMSUNG SpinPoint F2 EG (HD154UI)

md11 (500GB) =
SAMSUNG SpinPoint M7E (AF, HM641JI)
+
TOSHIBA MK5056GSY
Here I loosing some capacity, but I dont have any another 640GB drive :frowning: .

From LVM perspective it looks like:

  PV /dev/md7    VG default         lvm2 [<3,64 TiB / 2,44 TiB free]
  PV /dev/md10   VG default         lvm2 [465,63 GiB / <17,68 GiB free]
  PV /dev/md3    VG default         lvm2 [465,63 GiB / 0    free]
  PV /dev/md1    VG default         lvm2 [931,38 GiB / 0    free]
  PV /dev/md5    VG default         lvm2 [<5,46 TiB / <5,46 TiB free]
  PV /dev/md6    VG default         lvm2 [1,36 TiB / 0    free]
  PV /dev/md8    VG default         lvm2 [465,63 GiB / 0    free]
  PV /dev/md2    VG default         lvm2 [14,55 TiB / 0    free]
  PV /dev/md11   VG default         lvm2 [465,63 GiB / 0    free]
  PV /dev/md9    VG default         lvm2 [<7,28 TiB / <7,28 TiB free]
  PV /dev/md4    VG default         lvm2 [465,63 GiB / 0    free]
  Total: 11 [35,47 TiB] / in use: 11 [35,47 TiB] / in no VG: 0 [0   ]

It looks like “buying in the scrapyard”, but thanks to redudancy, I am fine. Many of my drives are 2.5" from notebooks, where I upgrade them to SSD and owner give me old drive as tradeoff upgrading and migration.

OK, this is a little offtopic :grin: . I would like to just show a different setup for one big node living actually 3+ years :slight_smile: .

1 Like