question : why truenas SCALE still can't scale a vdev by just adding disks to it ???

[Jeff63]

I've choosen Scale because its doc said that it could be scaled in or out easily, so I was thinking it was possible to create a zvol/vdev with 6 disks at fist and then later expand it to 10 or more disks when needs would be there.

reading this my expectations was that I will still have ONE RaidZ2 zvol with more disks and so more space in it for the datasets.

btw after adding that 4 new disks to it what I see is TWO RaidZ2 ! one with the original 6 disks and a new one with the same name with the 4 new disks...
the first one is still full, and the new is empty.

that's not exactly what I'll call 'scale'd storage.

what the heck ? if I need to backup and kill all each time I need to add disks to expand the storage of that nas what's the point to use such stuff that horribly add work on top of already too much work to do ??? where is the Scalabilty announced ?

 

[chuck32]

if I need to backup and kill all each time I need to add disks to expand the storage of that nas what's the point to use such stuff that horribly add work on top of already too much work to do ??? where is the Scalabilty announced ?

If that's a regular occurrence for you, you're not properly planning ahead.
Currently you cannot expand an existing RAIDZ2 VDEV, i.e. changing a 6 wide RAIDZ2 to 10 wide.

Just to clarify, did you actually extend your current pool or did you add a new one? I personally haven't played around with it, but I thought in order to extend a 6 wide RAIDZ2 pool you would need to add another 6 wide RAIDZ2 Vdev.

Post the output of zpool status

 

[William Grzybowski]

There is a feature in the works for this in upstream OpenZFS, should make an appearance in 2.3. TrueNAS will eventually grow this ability (possibly in ElectricEel) but for now its as @chuck32 explained.

 

[Etorix]

reading this my expectations was that I will still have ONE RaidZ2 zvol with more disks and so more space in it for the datasets.

Obviously you have not done your homework of properly reading how ZFS works, and you do not quite grasp data centre lingo ("scale out" means "across multiple servers"… and this functionality is not yet in SCALE).

ZFS is very inflexible for raidz layouts, and you have to design properly beforehand. Expanding one drive at a time is not an option for raidz—ZFS rather expects expanding one disk shelf at a time. If you go for mirrors, you may expand by a pair of drives at a time.

The 'hidden' cost of using ZFS for your home NAS

It's perfectly fine to conclude that ZFS is not for you, and rather look at Unraid or OpenMediaVault for the flexibility you want.

Note that, while raidz expansion is "planned", it has been so for years and whenever it eventually lands it will come with many caveats. In particular, previous data will NOT be reflown according to the expanded layout, wasting padding space; this implies that repeatedly adding one drive at a time will NOT play out well for available space.