Likelihood of data failure/advice

[Cam Hauver-Reeves]

So here's the deal. I'm basically the furthest thing from an advanced user you can imagine. I built my first FreeNAS sever a few months ago, with no basic understanding of ZFS (I didn't even know what ZFS was).

Despite my lack of knowledge, I somehow got my sever up and running. I use it for Plex, backups from a Windows 10 laptop, backups from a Mac, and RAW photo storage.

I currently have one 3TB drive, but it's nearly out of space. My plan is to add another VDev with a new single 3TB drive, then create a Zpool with both of the VDevs. I probably won't ever need to add another drive once it's up to 6TB, if anything that'll become excessive in my future. So my question is - when dealing with smaller sizes like this, do I need to worry as much about failure of a VDev and losing data? Do I even have another option if I don't want to lose all the data I currently have?

Thanks, Cameron

Hardware:
AMD Athlon X4 760k
Gigabyte Motherboard (can't remember exact model, but it's a consumer one for a gaming PC, not server grade)
8GB DDR3 Kingston Hyperx Fury
3TB WD Red


Sent from my LG-H918 using Tapatalk

 

[DrKK]

Sir:

Single-drive vdevs are something that *NONE* of us recommend for any data. A ZFS pool dies if, and only if, *any* vdev within the pool dies. In your case, that means you can tolerate *ZERO* drive failures. When a vdev fails, you (generally) lose everything.

I strongly recommend that you figure out some way to use vdev's that EACH contain multiple drives. A ZFS pool with no redundancy is an accident waiting to happen---an accident you are very unlikely to be able to recover from. We just want to keep your data safe, sir.

If you're after 6TB of storage, the MINIMUM I would suggest is a single vdev, containing 3x3TB drives, in RAID-Z configuration.

 

[rs225]

The other way is to make a backup or use replication to another system/pool.

 

[Cam Hauver-Reeves]

Sir:

Single-drive vdevs are something that *NONE* of us recommend for any data. A ZFS pool dies if, and only if, *any* vdev within the pool dies. In your case, that means you can tolerate *ZERO* drive failures. When a vdev fails, you (generally) lose everything.

I strongly recommend that you figure out some way to use vdev's that EACH contain multiple drives. A ZFS pool with no redundancy is an accident waiting to happen---an accident you are very unlikely to be able to recover from. We just want to keep your data safe, sir.

If you're after 6TB of storage, the MINIMUM I would suggest is a single vdev, containing 3x3TB drives, in RAID-Z configuration.

I may be able to trim down my backups and files to not need a new drive, but I had a thought for redundancy with the current setup.

Okay here's my idea now:

Buy two new 3TB WD Reds, create a vdev with full mirroring (I can't remember which RAID it would be, but resulting in keeping the 3TB total capacity but have full redundancy). Then, use drive replication to replicate the data from my current 3TB RED to one of the new ones. Then I'd probably sell the old one after all the data is on the new one.

Would this work? Do you have any ideas for alternatives? Thanks

Sent from my LG-H918 using Tapatalk

 

[Robert Trevellyan]

create a vdev with full mirroring (I can't remember which RAID it would be

That would be a mirror ;)

Would this work?

It would leave you with a 3TB raw storage capacity, and redundancy, which is 100% better than what you have now.

You should still have a backup strategy for any data you care about.

 

[Cam Hauver-Reeves]

That would be a mirror ;)

It would leave you with a 3TB raw storage capacity, and redundancy, which is 100% better than what you have now.

You should still have a backup strategy for any data you care about.

Great, thanks. Do you have any recommendations for a backup strategy? I've heard good things about Crashplan. Or were you referring to a hardware backup?

Sent from my LG-H918 using Tapatalk

 

[Robert Trevellyan]

Do you have any recommendations for a backup strategy?

Your options are almost unlimited. They range from replication to another pool within the same box, all the way to replication to an offsite FreeNAS or cloud storage. If you spend some time thinking about which data you care about, and what kinds of risks you want to mitigate (user error, hardware failure, fire/theft etc.), then do some basic research, you'll be able to ask more specific questions.

I consider CrashPlan a good solution for client system backup, but it requires quite a bit of hoop-jumping to backup from FreeNAS.

I use HashBackup to send important data to BackBlaze b2, but that requires a level of comfort with command-line usage and CRON job scheduling.

EDIT: <hint> there are lots of threads about backup in these forums. </hint>