[SOLVED] Proxmox 8.2.4 wont allow me to create a raid0 pool.

[Avery]

Let me start out by saying this is a home setup, nothing is going on this pool that I don't have saved elsewhere. So if we can skip all the you will loose everything bit that would be great. I have two new 20tb drives I shucked from WD external cases. However I will be adding two more drives as funds allow so I can create a raid0-1 pool later or raid 5? We shall see. I haven't messed with raid much in probably 10 years and Ive learned a lot has changes. I am a noob when it comes to stuff like this however. The reasoning for needing raid0 is speed, I have been running jellyfin on a pc for awhile and with having lots of devices my house we can exceed the read speed of one drive, and currently between movies,shows,family photos were pushing 17TB and growing. So I need speed and space.

So I tried to create this pool in the GUI and learned that the only way to do that is to use the cli.
So I went to the shell of my pve node and gave this command: (I'm just copy and pasting everything)

root@pve:~# # zpool create deadpool ata-WDC_WD200EDGZ-11B9PA0_2GGB88ZV ata-WDC_WD200EDGZ-11B9PA0_2HG07DVN
root@pve:~#
I get no errors or anything, it just ignores it. Ive also tried this
root@pve:~# # zpool create -f -o ashift=12 zpool /dev/sdb /dev/sdc
root@pve:~#
Ive also tried this
root@pve:~# # zpool create -f -o ashift=12 <deadpool> </dev/sdb> </dev/sdc>
root@pve:~#
When i run zpool status i only see this other pool

root@pve:~# zpool status
pool: pool
state: DEGRADED
status: One or more devices has experienced an error resulting in data
corruption. Applications may be affected.
action: Restore the file in question if possible. Otherwise restore the
entire pool from backup.
see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-8A
config:

NAME STATE READ WRITE CKSUM
pool DEGRADED 0 0 0
ata-ST18000NM000J-2TV103_ZR5CQECK DEGRADED 0 0 32 too many errors

errors: 8 data errors, use '-v' for a list
root@pve:~#
Ive tried rebooting but that doesn't seem to make much difference. Any suggestions would be appreciated
Thanks

 

[esi_y]

a raid0-1 pool later or raid 5?

RAIDZs (equivalent to RAID5 or 6) are a bad idea.

The reasoning for needing raid0 is speed

Not with ZFS, not without SSDs for a caching device, at the least.

can exceed the read speed of one drive

What is your networking? Modern spinning drives get 200MB/s sequential and regular 1Gbps LAN gives you ~ 100MB/s in total.

---

The command you are issuing is documented [1], but I don't want to post you an example blindly so as you do not cause more trouble to your existing pool. You would preferably want to refer to your vdevs (physical drives) by /dev/disk/by-id/... paths. But also, I would prefer to create a GPT partition [2] over each first and refer to those as vdevs instead.

I think you better off research more on ZFS first or skip this in favour of e.g. TrueNAS or similar solution. E.g. you can't do certain things with pools without destroying them and recreating from scratch:

I will be adding two more drives as funds allow so I can create a raid0-1 pool later or raid 5? We shall see.

[1] https://openzfs.github.io/openzfs-docs/man/master/8/zpool-create.8.html
[2] https://fedoramagazine.org/managing-partitions-with-sgdisk/

 

[UdoB]

root@pve:~# # zpool create deadpool ata-WDC_WD200EDGZ-11B9PA0_2GGB88ZV ata-WDC_WD200EDGZ-11B9PA0_2HG07DVN
root@pve:~#
I get no errors or anything, it just ignores it. Ive also tried this
root@pve:~# # zpool create -f -o ashift=12 zpool /dev/sdb /dev/sdc

Stupid question: the second "#" is not really there - or is it? It would make your whole "command" a comment!

 

[Avery]

Oh my I'm an idiot.... Thank you all for getting back to me. The second # was the problem. The first one was already on the cli and I though I had to type another. And second it was my network speed I was saturating. I've been working on and building computers most my life, I'm not sure how I over looked that. I will admit I suck with cli and linux. I am not concerned with my other pool tbh, as I said I have backups and backups of those backups. Honestly I am not sure if this is what I should be using or not, It all started simple with me setting up opnsense on a small 1U server because my netgear router became unstable and it just kinda snowballed. I havent had a great few years and with my adhd (Im 40 btw) I needed a distraction so I built this 4U unit that I have a vm of opnsense, jellyfin,cockpit and windows 10. Im sure by the time I know what im doing I will have rebuilt it a few times over. Also I like to tinker and learn new things. When I first started playing with raid I think the only options were raid0 0-1 or jbod. Would yall have built this differently? Thanks again!

 

[Kingneutron]

So with the "raid0" pool you have, down the road you can attach a same-sized drive to each disk, (let resilver finish before doing the 2nd one) and end up with a "raid10 equivalent" - aka a pool of mirrors. Then you will have redundancy + speed.

https://github.com/kneutron/ansitest/blob/master/ZFS/mkdynamic-grow-pool-boojum.sh
^ Start at line 163

 

[esi_y]

I will admit I suck with cli and linux. I am not concerned with my other pool tbh, as I said I have backups and backups of those backups.

That's alright, just be aware that e.g. commands like zfs destroy do not even ask for confirmation.

 

[Avery]

I just wanted to thank you guys once more for all the help, and advice. I bookmarked all those links esi_y and Kingneutron, lots of good info in them.