New To ZFS and TruNas. Simple Question

TrueNAS General
, ,
1

ElectricEel-24.10.0
AMD Ryzen 5
64gb ram

I have browsed the forum and haven’t seen these answered.
Question 1.
In past RAID setups, I have hade parity drives and data drives, in ZFS there is no Identifiable Parity drive(s). Am I ok to assume that ZFS runs similar to Raid6 in that the parity is spread among the disks? (I currently have a RAIDZ2 |6 wide)

Question 2.
In TruNas, in general, is there a place to view logs for completed task results for things such as the SMART testing, Scrub, etc. I schedule them to run under the Data Protection tab yet i am unsure of where or if I can view reports/logs of their progress.

Thank you for any help.

2

Sounds like Unraid. ZFS does not do that.

Data and parity is striped across a vdev (RAIDZn), yes.

There’s a tasks icon in the top right of the dashboard.

3

@ winnielinnie Thank you for the Task Info, I appreciate your help.

The RAID setup I was referring to was Hardware RAID (Not sure if I have to caps this or not), this is my first venture into software RAID and ZFS in general. I tried Unraid software and just liked the TruNas software better.

So If I am understanding, all ZFS(n) setups, include the parity across all drives. That must be why it is becoming so popular.

I’m really excited to learn on this new setup.

Again, Thank you…

4

ZFS pools are comprised from vdevs, which are constructed by member devices (drives).

If a vdev is a single stripe, it’s a non-redudnant vdev. You lose it, you lose the pool.

If a vdev is a mirror, it has redundancy, but no “parity” is involved.

If a vdev is RAIDZn, it has redundancy, and stripes the data and parity across all members.

A pool can be made of multiple vdevs.

Do not use hardware RAID with ZFS. Make sure you build your vdevs with direct access to the drives themselves.

5

I think that user was referring to RAID-3, which uses dedicated disk / column for parity. RAID-5/6 uses spread out parity, similar to RAID-Zx.

To be clear, RAID-Zx does not have to use full width writes, like RAID-5/6 does… Thus, “similar” for spreading out parity, but not “same”.

2 Likes
6

A simplified explanation…

With ZFS, each record can be only as big as the width of the vDev (including parity), but it can be smaller. So on a 5x RAIDZ2 i.e. 3 data, 2 parity, a record can be between 3 and 5 wide i.e. between 1 and 3 data blocks and always 2 parity blocks. These are then written to separate drives.

2 Likes
7

That’s actually a very unusual arrangement (apart from things like Unraid). It does exist in RAID 3 or 4, but those RAID levels are very rarely used; 5 and 6 are by far the most common parity RAID levels. And as already said, in ZFS parity RAID (RAIDZn), the parity is spread across the disks in the vdev.

1 Like
8

Anyway, yes, RaidZ2 is effectively dual drive redundancy. Similar to RAID6.

Where it gets tricky is that you can have multiple raidz2 virtual devices (vdevs) striped together in a pool, which is similar to Raid60.

IMO ZFS is popular because of its features and safety. Ie everything is checksummed and it supports snapshots, replication, compression, block cloning, etc.

9

I started in the industry using Punch card systems and 12" platter drives. Then thought I had the whole new car feel when hardware RAID came out. You could only imagine my excitement when RAID became integrated and extended for more failsafe use. Now, I feel like a chimp with its first cellphone as I learn to design, setup, and put TruNas into usage.

I truly appreciate everyone’s help and patience as I come out of the tall grass and get to enjoy the fresh new lawn, I’ve heard so much about.

Thank you,
:monkey_face: :iphone:

1 Like
10

Punched cards? You were lucky.

When I as a lad it was paper tape - and if you were lucky you got to keep the tiny circles of paper and use them to keep warm in t’winter.

3 Likes
11

Four Yorkshiremen- Monty Python on Make a GIF

2 Likes