Jump to content

Planning migrating to unRAID


Hanfufu

Recommended Posts

Hello all,

 

I have been wanting to migrate my Plex server from WS2016 to unRAID for quite a while now, and i finally have the possibility of buying 3x14TB drives, of witch 2 are meant for parity.

 

I have however, seen some concerning posts regarding speeds on esp writes. Also i seem to recall something about that the more disks that you add, the slower it can get, but is there some truth to that? 

My server has a Asrock x570 board, 3950x, 48GB RAM and support, and my disk setup would be this:

 

2x14TB for parity, and for data:

1x14TB Seagate enterprise.

5x10TB IronWolf disks,

2x6TB Normal disks,

1x12TB WD something

 

And i was thinking of doing a RAID 0 with 4x2TB Kingston NVMe disks for speed and redundancy (using an Asus HyperX PCIe->4xm.2 adapter), and use a 1TB NVME m.2 disk for cache.

 

Anyone have any idea what i could reasonably expect, performance wise from this setup?

If reads/writes are as low as like 40MB/secs as i have heard, it will take literally forever to transfer my existing data (located across all the non 14TB disks mentioned), as it will saturate the cache very fast.

 

And on another note, is GPU passthrough. I went from Xeon CPUs to that ryzen one, and had never  before worried about PCIe lanes, but i apparently have to with the only 20 available on ryzen CPUs - I have read that it IS possible to pass through a GPU to a VM, even though its the only one in the system, that can be "released" by unRAID when bootup is completed, is that correct?

 

There is no iGPU on the CPU, and with the Asus pcie x16 card to m.2, i will only be able to have 1 GPU, currently a 1050TI that will have to make do with a x4 interface so it is not possible to have multiple GPUs in the system.

 

I hope someone has something to share on these things, and i wish everyone a nice evening :)

Link to comment
11 minutes ago, Hanfufu said:

writes are as low as like 40MB/secs as i have heard

Writes to the parity array are slower than single disk speed due to realtime parity updates. If you transfer your files before adding parity disks then you would get single disk speed. No point in caching initial data load since it would fill up.

Link to comment

It is good that you plan for your PCIe lanes

 

1 hour ago, Hanfufu said:

And on another note, is GPU passthrough. I went from Xeon CPUs to that ryzen one, and had never  before worried about PCIe lanes, but i apparently have to with the only 20 available on ryzen CPUs - I have read that it IS possible to pass through a GPU to a VM, even though its the only one in the system, that can be "released" by unRAID when bootup is completed, is that correct?

 

There is no iGPU on the CPU, and with the Asus pcie x16 card to m.2, i will only be able to have 1 GPU, currently a 1050TI that will have to make do with a x4 interface so it is not possible to have multiple GPUs in the system.

 

But unless I missed it, I do not see any mention of a plug-in SATA or SAS controller. X570 boards generally come with 4 to 6 SATA ports, maybe 8 if you are lucky ?

You list 11 HDDs, so you will need PCIe x4 or x8 for that too. Maybe electrical x2 ?

 

 

Link to comment
30 minutes ago, ChatNoir said:

It is good that you plan for your PCIe lanes

 

 

But unless I missed it, I do not see any mention of a plug-in SATA or SAS controller. X570 boards generally come with 4 to 6 SATA ports, maybe 8 if you are lucky ?

You list 11 HDDs, so you will need PCIe x4 or x8 for that too. Maybe electrical x2 ?

 

 

Thank you for the reply, and well spotted, I forgot to mention i have a pci-e cheap noname sata controller with x4 ports. 

I still am unsure how the lanes work, as there are 3 small x4 slots on the Motherboard, besides the 2x x16 physical (of which one operates in x4 and the other x16 or x4/x4/x4/x4 bifurcation mode).

I seem to recall something about some lanes wired to the chipset, and was thinking that was the case with the small slots. Otherwise you would have 0 use for more than 2 slots in a Ryzen gaming pc (x16 and 1 x4 tales up all 20 lanes wired to the CPU), and the motherboard is more or less gaming oriented.

 

But maybe im wrong, and would love to be corrected :)

Edited by Hanfufu
Link to comment
1 hour ago, trurl said:

Writes to the parity array are slower than single disk speed due to realtime parity updates. If you transfer your files before adding parity disks then you would get single disk speed. No point in caching initial data load since it would fill up.

Interesting, so its possible to add all drives to a pool (or what its called), copy data to the volume with single disk speeds, and then add 2xparity drives without any problems, when all data has been copied?

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...