How to change case sensitivity

[thany]

I'm trying to set up my pools. It makes sense for them to be case-insensitive in my scenario. However, I cannot figure out how to do that.

When I set up a pool, I'm not asked for any options basically, only the name and which disks to include. So any options can only ever be set after the fact. That's okay of course, were it not that the `Case Sensitivity` option is theonly one greyed out, but it doesn't say why. The little ?-icon only vaguely explains what the options mean, but that doesn't help me selecting one.

Please advice - how do I set a pool to be case-insensitive?

 

[anodos]

You create a dataset and make it case-insensitive during creation. This is an immutable property.

 

[thany]

I'm not familiar with datasets. How does that relate to a pool?

Either way, so I create a dataset inside this pool? But then what is the option for Case Sensitivity doing in the pool options if it only every applies to a dataset? Why is it there?

 

[anodos]

/scale/scaleuireference/storage/

 

[thany]

That only describes the screens, and vaguely what a pool is. Sorry but I can't find much coherency in there, nevermind the relation between these types of things that I can (and should, inexplicably) create.

 

[sretalla]

Please read the ZFS primer.

https://www.truenas.com/community/resources/introduction-to-zfs.111/download

 

[thany]

Still no answer to this:

> But then what is the option for Case Sensitivity doing in the pool options if it only ever applies to a dataset? Why is it there?

 

[anodos]

Still no answer to this:

> But then what is the option for Case Sensitivity doing in the pool options if it only ever applies to a dataset? Why is it there?

You were linked to documentation so that in theory you could read it and find your answers.

Code:

root@TN3[~]# zpool get name dozer
NAME   PROPERTY  VALUE  SOURCE
dozer  name      dozer  -

this is my pool

Code:

root@TN3[~]# zfs get name dozer
NAME   PROPERTY  VALUE  SOURCE
dozer  name      dozer  -

This is the default root dataset for a pool.

When you're doing things where case-sensitivity is relevant you're operating on the dataset layer. If you're clicking around in the GUI settings, you're altering the settings for your pool's root dataset.

 

[thany]

In theory yes, but I'm asking a question. The answer is probably buried somehwere in the documentation, but really is it so much to ask?

You answer still doesn't explain why on a pool there is a case-sensitivity option that can never be set or changed. You're saying this is because a pool equals a root dataset. That's great, and I understand that. But then why is this option available (and disabled) on a root dataset? Why is it called a pool, and not a root dataset? Why is the option there *at all* if it can *never* be changed after the fact? Why make the UI such that it appears to be changeable under "some" condition, unknown what that condition is, instead of just displaying a static label?

 

[anodos]

But then why is this option available (and disabled) on a root dataset?

Because a system administrator may wish to know the status of properties even if they're immutable once the dataset is created.

The storage form is in the process of being redesigned in the 22.12. You can download the BETA and take it for a spin. If you have specific actionable feedback, you can feel free to file tickets on our bugtracker at jira.ixsystems.com.

 

[anodos]

Do note that if you want to create a case-insensitive share, you can simply create a new dataset (e.g. dozer/SHARE) and share out the path /mnt/dozer/SHARE. Generally speaking, it is not best-practice from an administrator's perspecitve to expose the entirety of a pool ("/mnt/dozer") an NFS export or SMB share.