Hard drives usage and disk space

[RobGoss]

I'm not sure that question is very clear... is your NAS also your home server? Is your home server running a unix or linux based Operating System?

To put it simply...

You can use the disks in a ZFS pool in another system that supports ZFS and you will be able to import and use your pool data. (a few caveats around the version of ZFS that you're using, but if you're using recent versions of OpenZFS, no problem generally)

Yes my Nas server is my home server. My home server is not using any form of Linux

 

[sretalla]

Yes my Nas server is my home server. My home server is not using any form of Linux

OK... your NAS / Home Server is running TrueNAS... ? (in which case it is running either Linux or Unix)

Then I'm back to not understanding the question... if there's a problem with your NAS, can you use the disks from your NAS in your NAS? am I getting the question right?

 

[RobGoss]

OK... your NAS / Home Server is running TrueNAS... ? (in which case it is running either Linux or Unix)

Then I'm back to not understanding the question... if there's a problem with your NAS, can you use the disks from your NAS in your NAS? am I getting the question right?

I misunderstood your reply sorry, The reason I wanted a home server was to continuously make backups of the content I produced and be able to access it easily if something went wrong. I do understand if a disk fails I may not be able to recover any data.... You are correct I assume Nas is Linux base. I ran Ubuntu for many years on my order machines

 

[sretalla]

You're running SCALE, so it's Debian Linux.

I'm still not clear on the question:

if something were to happen to my server / Nas would I be able to access that data on the hard drives I have on my home server?

But if I assume you want to know if you'll be able to use the disks elsewhere, my answer from before covers it:

You can use the disks in a ZFS pool in another system that supports ZFS and you will be able to import and use your pool data. (a few caveats around the version of ZFS that you're using, but if you're using recent versions of OpenZFS, no problem generally)

 

[RobGoss]

Yes (simple answer).

To explain fully, a ZFS pool uses VDEVs (at least one) to provide the storage. Since the VDEV(s) are joined together in a stripe, if any one VDEV is lost, the pool won't have the required information to continue operating and all content will be unavailable.

So what's recommended in most cases is to use redundancy within a VDEV to ensure that a single disk loss (which is one of the more common things to happen to a spinning HDD, although relatively rare during the first 4 years or so of a disk's life) doesn't render the pool unavailable.

Redundancy can take the form of Mirrors (1 disk for data and additional disk(s) for redundancy), RAIDZ1 (3 or more disks where 1 disk worth of capacity is given to storing "parity" - sufficient information to rebuild the block based on all remaining disks if one is missing) RAIDZ2 (4 or more disks where 2 disks capacity are given to parity) or RAIDZ3... 5 or more disks, 3 to parity.

In each of those cases, losing 1 disk will result in the pool continuing to operate with full access to read and write files as normal (although the pool will show the administrator that it's in "degraded condition" due to the missing disk).

Obviously there are positives and negatives to each type of redundancy, which can be weighed in terms of cost of the parity versus performance and data security...

Mirrors are costly in that you need to buy 2 disks to have 1 disk worth of capacity. But for that, you get good performance (particularly in terms of write IOPS). You run the risk of a failure of the remaining disk or data corruption when one fails until you are able to add a replacement disk, but that risk is reduced due to the relatively low load of "resilvering" (copying/reconstructing back the content to the replacement disk) as you're just copying blocks 1-1.

RAIDZ1 costs you 1 disk out of your VDEV in terms of capacity, so it's not too bad if you have 5 disks (you still get the capacity of 4). Throughput for writes can be probematic for large amounts of smaller files and for things like Virtual machine disks or iSCSI as the IOPS of a RAIDZ1 VDEV is about the same as that of only 1 of the member disks (about 1-300 IOPS for most HDDs). With a failed disk, you can continue operating, but the resilvering process is hard on the remaining disks, increasing the likelihood of a second failure and/or corruption of data before you can get back to normal with the replacement. For that reason, it's not encouraged to use RAIDZ1 with disks larger than 1TB (you could stretch it to 2TB).

RAIDZ2 is more of the same as RAIDZ1, but with the difference that with one disk lost, you are still protected against corruption and even against a further disk failure during the resilver process. Highly recommended if data integrity and security is a priority.

RAIDZ3 is even more paranoid, allowing for the pool to operate with 3 disks missing and providing protection even in the case of 2 missing.

Generally, if a pool is to have more than one VDEV, you would try to match the sizes and VDEV types in order to ensure you're not taking risks for the whole pool that you hadn't decided to take for one of the VDEVs.

Anyway, that's probably information overload for you at this point, so I'll suggest that you look over in the Resources section in the Fundamentals category for more reading material if you're so inclined.

Thank you so for this detailed post I will surely make screen shots of it so I can refer to it when needed