Thanks everyone for helping me with this problem.
I decided to move all data to another pool (no SMR drive) and remove those drives from the server.
I will mark this as Solved.
Just as a bad designed example: one single pool, RaidZ2, three devices for the Special Device because two were not stable - and one being attached by USB. Again: bad design.
What I can see here is this:
root@pvec:~# zpool iostat -v 5...
Do you expect a connection error and not the probably very old 2TB WD red drive?
VRPF failed, right?
You pulled it out and inserted the 9TN2, right?
BTW don't do that if you have enough bays. Just insert the new drive and let the resilver do...
PBS does really, really do random writes. It needs IOPS - as many as it can get.
For ZFS there is the rising problem of fragmentation of the free space --> it is not guaranteed that a single 2-4 MB chunk can be written sequentially at one...
Not what I am seeing in the real world.
Let's start with a fresh VM where nothing is already backed up:
zpool iostat poolBackup 1
capacity operations bandwidth
pool alloc free read write read write...
To be more precise: L2ARC is not optimized for read workloads. It is an added read cache that caches stuff that evicted ARC before. It does only do read cache, never write cache.
Not sure if I would agree with that statement.
I have not run...
Of course not :) They are bottom of the barrel trash SSDs! Look at the price.
I use it as a boot disk. Which is mirrored with another drive. When using a drive, I suspect a firmware issue leading to total failure and plan my vdevs accordingly...
run it again but this time with a PBS destination so you get a TLS result on PVE.
Then run the same benchmark again on PBS, where you don't get a TLS result.
That is why everybody will recommend you to use mirrors for blockstorage hypervisors and not RAIDZ.
RAIDZ can be good for files, but is not good for block storage.
Short overview on how this works:
4k secotor size is the default. So you will...
PBS is a rather poor technology for file backups. If you don't need blockstorage but files, something like robocopy, rsycn, zfs send or the likes will perform miles better, because it only has to read metadata (and for ZFS not even that). PBS on...
Jesus, these prices are insane.
Anyway, the PSU was used (Privatanbieter means private seller) and you don't need a CPU cooler (AMD has one included), nor a case fan.
In regard to the discussion: I remember that @Falk R. mentioned several times in the German forum that using ZFS on HW RAID is ok if (and only if):
You know what you are doing (which imho doesn't hold true for people calling others "dogmatic"...
The point is: you can not know that.
Of course you are free to use any available technology you want - and a lot of installations are using hardware Raid (edit: ... without a checksumming filesystem). No fight here; if it works for you/them...
I don't think is does 4k r/w and even on my consumer hardware I linked to you in another post I can easely reach 100MB/s.
Anyway, in the real world, PBS will be way less relevant, since you only need to send changes. So in your case the local...
I've already written what it does, and am not about to go on the full defensive. Reiterating succinctly: This is a wrapper for PBC. It adds inheritable backup config from within the dataset properties (vs. remembering to modify your PBC scripts)...