I finally ran out of disk space on my PLEX server.

[genBTC]

Bruh, it says 11.9 GIGS. you have to look on the TOP line, for the maximum space. 6.9TB terabytes.

 

[Chris Moore]

I mean, I'm not hurting for money...My plan was to buy another 8TB External HD and just xfer all my movies from the server onto my externals. Then I was gonna buy 3 more WD 4TB Red drives and then re-do them all together. But seeing how I only have 2, I would just buy 2 more WD Red drives and re-do the free NAS

Your plan to move everything off the storage, rebuild it and then bring the data back would appear to be your only good option. If you do it right, you might even be able to preserve your PLEX watched list and other settings. Please do ask for advice on that also, Good luck.

I would suggest that you shoot for a set of 6 drives in RAIDz2. That would give you 14TB of storage in place of the 6 (ish) that you have now and it would allow for a drive failure without loss of data. Depending on the exact nature of your hardware, you might be able to do this with only adding drives to your existing setup and a some reconfiguration of the software and shuffling of data.

 

[therobmyers]

Thanks for all the advice. As far as the plex watched list and stuff I don't mind losing all of that. I think that's what I'll do then.

 

[Stux]

Your plan to obtain an 8TB external is a good one. Once everything is done you can use it as a nice backup device for your most important content.

You currently have 2 4TB drives with no redundancy.

We like to recommend dual disk redundancy, as there is a very real chance that if a single disk fails, that you may lose all your data if another disk even has a partial failure during the rebuild (which is very stressfull to disks), if you only have single-disk redundancy.

Now, RaidZ2 is a good way of providing this dual disk redunancy. The way it works is you end up with n-2 * the size of your disks for storage. So 4 disks of 4 TB gives you 8TB, and 6 disks will give you 16 TB (ie (6-2) * 4TB).

So, a really good way to go would be to backup your < 8TB of data, and then make a nice 4-6 way RaidZ2 for 8 to 16TB of storage with dual disk redundancy, depending on how many extra disks you have. Restore your 8TB external to your new pool, and then you can use the 8TB external for backup purposes.

If you wanted to add redundancy to your current pool, you could mirror each drive... and that would give you full redudancy for each drive... but no extra space. Meaning to get extra space you would still need to add an extra drive or two.

In which case, the best thing to do is to go with the 6-way RaidZ2 anyway.

Alternatively, you could purchase two 8TB drives. And then make a pool out of 2x4TB and 2x8TB and end up with 12TB. And no backup device.

Assuming you want redundancy.

And the more disks you have, the more chance of one of them developing a fault and taking your data with it... if you don't have any redundancy.

Hope this helps :)

 

[Chris Moore]

Now, RaidZ2 is a good way of providing this dual disk redunancy. The way it works is you end up with n-2 * the size of your disks for storage. So 4 disks of 4 TB gives you 8TB, and 6 disks will give you 16 TB (ie (6-2) * 4TB).

That is a terribly inaccurate way to calculate actual available (usable) pool space. Here is a calculator that is actually very accurate, even on the 60 drive pool I use at work, it is only off by a small percentage.
http://wintelguy.com/zfs-calc.pl

 

[Stux]

That is a terribly inaccurate way to calculate actual available (usable) pool space. Here is a calculator that is actually very accurate, even on the 60 drive pool I use at work, it is only off by a small percentage.
http://wintelguy.com/zfs-calc.pl

It doesn't take into account overhead, and slop space allocations, but other than that, it is a perfectly reasonable way to estimate storage after redundancy.

If 500GB out of 16TB is an issue for your capacity planning, you have bigger problems.

 

[Chris Moore]

If 500GB out of 16TB is an issue for your capacity planning, you have bigger problems.

It isn't my planning. I have my hardware built already.

The difference, if you take a turn using the calculator I linked to, is more like 2TB, not 500GB.

Don't take it personally, but capacity planning makes a lot of difference. I have a pool that I manage for work that has over 260TB of data stored. If we didn't plan that properly, the data we needed to store would not have fit in the pool. The bigger the data, the more important the planning.

 

[Stux]

The difference, if you take a turn using the calculator I linked to, is more like 2TB, not 500GB.

That calculator calculates 14.06TiB usable storage (before reserving 20%) for a 6-way RaidZ made out of 4000GB disks.

14.06TiB is about 15.46TB, vs the estimate of 16TB. The 540GB missing is mostly the 1/32nd that is reserved as slop space.

Slop space being the last 3.2% or so of the drive that ZFS will only use if it *really* has to (ie to allow file deletion).

To the OP. You should plan to upgrade your storage when you get to about 80% usage, and you should have implemented that plan when you get to about 90% usage.

 

[danb35]

The one piece of good news for OP is that he can, in fact, add that third drive without losing redundancy, as he doesn't have any redundancy to lose in the first place. And in that case, it can work exactly as he describes in his OP--he can extend the pool with another disk, the space will be immediately added and available to the pool, and he can continue filling it up.

The obvious problem (which is already discussed here, and I noted above) is that there's no redundancy, so when any disk fails, all the data goes away.