New to drivepool, good use case

[xelu01]

New to drive pool and wanted to get some help on if this is the right use case before I purchase it. I have a truenas machine at home with all my data in a raidz1. I want to have a second set of this data in case anything happens. I have a many mixed size drives to use and was going to use them with windows storage spaces but heard about drive pool in the past and thought I'll drop by to see if anyone can help me with this. I want to just use this as a backup of my truenas data but didn't want to have a second truenas machine. Can I do this with just having a pool of my drives without duplication, since it's a duplicate of my data? Or would you still recommend duplication? The only worry is if both machines go down, but if one does then i got a second copy of my data.

[Shane]

Hi xelu01. DrivePool is great for pooling a bunch of mixed size drives; I wouldn't use windows storage spaces for that (not that I'm a fan of wss in general; have not had good experiences).

As for duplication it's a matter of how comfortable/secure you feel. One thing to keep in mind with backups is that if you only have one backup, then if/when your primary or your backup is lost you don't have any backups until you get things going again. Truenas's raidz1 means your primary can survive one disk failure, but backups are also meant to cover for primary deletions (accidental or otherwise) so I'd be inclined to have duplication turned on, at least for anything I was truly worried about (you can also set duplication on/off for individual folders). YMMV.

Regarding the backups themselves, if you're planning to back up your nas as one big file then do note that DrivePool can't create a file that's larger than the largest free space of its individual volumes (or in the case of duplication, the two largest free spaces) since unlike raidz1 the data is not striped (q.v. this thread on using raid vs pool vs both). E.g. if you had a pool with 20TB free in total but the largest free space of any given volume in it was 5TB then you couldn't copy a file larger than 5TB to the pool.

[xelu01]

2 hours ago, Shane said:

Hi xelu01. DrivePool is great for pooling a bunch of mixed size drives; I wouldn't use windows storage spaces for that (not that I'm a fan of wss in general; have not had good experiences).

As for duplication it's a matter of how comfortable/secure you feel. One thing to keep in mind with backups is that if you only have one backup, then if/when your primary or your backup is lost you don't have any backups until you get things going again. Truenas's raidz1 means your primary can survive one disk failure, but backups are also meant to cover for primary deletions (accidental or otherwise) so I'd be inclined to have duplication turned on, at least for anything I was truly worried about (you can also set duplication on/off for individual folders). YMMV.

Regarding the backups themselves, if you're planning to back up your nas as one big file then do note that DrivePool can't create a file that's larger than the largest free space of its individual volumes (or in the case of duplication, the two largest free spaces) since unlike raidz1 the data is not striped (q.v. this thread on using raid vs pool vs both). E.g. if you had a pool with 20TB free in total but the largest free space of any given volume in it was 5TB then you couldn't copy a file larger than 5TB to the pool.

Thanks, this is what I wanted then. I need to mix drives and slowly add more storage as needed.

For duplication, is it like raid 1 or does it work more like mirrored where it makes an exact copy of each file?

The likelihood of both going down at the same time is low but i understand what you're saying because if both go then i'm screwed lol.

My only thing about not duplicating is I heard the scanner would move files off the dying drive to the other drives if it has space, is this true? Would it be good to go to that and the cloud as well to make sure nothing happens? Like get it to move automatically my data if a drive is dying or get notifications if they are. Will the move on their own the data on the bad drive, or would I need to initiate it?

For bitrot, drive pool can't do much but the zfs copies should be taken care of and when copied over it should be the same, right?

 

[Shane]

"For duplication, is it like raid 1 or does it work more like mirrored where it makes an exact copy of each file?"

DrivePool's duplication is basically mirroring but file-based rather than block-based; if you set (or change) the duplication on a pool or folder to N copies then DrivePool will try to maintain N identical copies of the pool's or folder's files across the drives in the pool, one copy per drive. E.g. if you have a pool of five drives and you set pool duplication to x3, then DrivePool will keep one copy of each file on each of any three drives of the five drives in the pool. You have the option of real-time or nightly duplication (the latter will also run as a safeguard even if you've selected the former).

"My only thing about not duplicating is I heard the scanner would move files off the dying drive to the other drives if it has space, is this true? Would it be good to go to that and the cloud as well to make sure nothing happens? Like get it to move automatically my data if a drive is dying or get notifications if they are. Will the move on their own the data on the bad drive, or would I need to initiate it?"

StableBit DrivePool does have an option to evacuate drives if StableBit Scanner flags them as bad (e.g. via SMART warning); if it's enabled it happens automatically (unless it runs out of space, yes). Note that if a drive fails abruptly then it won't have time to evacuate the drive, which is where duplication is useful. DrivePool and Scanner can provide notifications about failed and failing drives respectively (including via email if configured). If a drive in a pool does fail, the pool will become read-only until the drive is manually replaced (or removed).

In a business environment, good practice for backups (IMO) is usually considered to be at least three copies of your files (the working copy and two backups) in at least three locations (your work machine, your onsite backup machine and another backup that's secured somewhere else). YMMV.