Heads up! 21.06 BETA out soon

[ornias]

What about a fix for those of us that have already moved from Core to Scale and are having ACL permission issues? Is there a script that be applied? I really just need to wipe out all ACLs and apply new ones.

You shouldn't have been moving to an ALPHA release, like said on the forums many-many times already.
So thats the biggest part of your problem at the moment... As it was pretty clear there is/was no clear migration path.

That leaves just wiping the ACL's and starting over really..

 

[cyrus104]

You shouldn't have been moving to an ALPHA release, like said on the forums many-many times already.
So thats the biggest part of your problem at the moment...

Hey man, you've done a lot at helping people out and I think you for that but I really think you are starting to give these forums a negative vib. I upgrade to alpha on a personal homelab system that doesn't have any data that I really care about (that's stored elsewhere). I've submitted questions regarding services and charts that I was having an issue with and have submitted Jira tickets as requested.

I've seen numerous posts where you berate people for asking for help or bringing up issues, some maybe rightfully so others not so much. Because I converted from Core to Scale on a plex storage to the alpha for testing, I feel my question is justified in seeing if there will be a fix or script as I move from 21.02 to 21.04 and soon on to 21.06.

Your knowledge has been great at helping out but your constant negative attitude toward most users is turning me off from coming here to learn and support, I'm turning to reddit users more than the official forum and spun up another storage solution to trying an avoid the negativity.

 

[ornias]

I upgrade to alpha on a personal homelab system that doesn't have any data that I really care about (that's stored elsewhere).

Okey, but that was not clear from your initial response.
As "moving to scale" never was supported in the first place.

I've submitted questions regarding services and charts that I was having an issue with and have submitted Jira tickets as requested.

I don't really have an idea what this is about tbh.

Because I converted from Core to Scale on a plex storage to the alpha for testing, I feel my question is justified in seeing if there will be a fix or script as I move from 21.02 to 21.04 and soon on to 21.06.

Well, migration wasn't supported at all (and isn't still) so there isn't really a magic pill.
It also isn't only an iX issue but Upstream issues as well that where blocking this.

Migrating these ACL's is quite complex and also one of the blockers for CORE-> SCALE migration at all. So I doubt there is such a solution and if there was, I'm not sure if it would be safe for data that is (lets say) 50% ex-CORE and 50% new-SCALE data. (hence some have ACL's that maybe or maybe not could be migrated, and some don't)

 

[cyrus104]

Thank you for the useful information, the migration that I did was just exporting and importing the zfs pools. I can now see that this upstream permissions issue is a complex issue due to it not just being with ZFS but the OSes that it's running on. While my homelab is decent, I do not (and I'm sure this applies to many others) have the spare space to migrate all of my data while I destory the zfs pool instead of exporting and importing it.

I don't have a personal grievance with you. I also know that you are "aware of your tone of voice" and from my perspective which means nothing here, the tone takes away the focus from the thread topic due to negative bias.

 

[Kasazn]

I have been testing alpha and nightly builds and I've managed to have SMB share is to do it on the Credentials > Local Users GUI. Only one folder can be set to one user at one time so it's a hassle for now.

 

[anodos]

What about a fix for those of us that have already moved from Core to Scale and are having ACL permission issues? Is there a script that be applied? I really just need to wipe out all ACLs and apply new ones.

There's not really a way to easily "undo" ACL changes. Options are the same as if you had chown -R <name>. The changes are written to disk, and you either have to (1) roll with it or (2) restore from a backup (or ZFS snapshot). The ACL editor for POSIX1E ACLs should be working in the WebUI soon. This portion of the webui is going through a major rewrite to make it overall _much_ more usable.

If you have a snapshot of the dataset from before the migration I do have an ACL restore tool that _might_ work. I wrote it for core and ported to scale, but it hasn't been tested yet.

 

[anodos]

I have been testing alpha and nightly builds and I've managed to have SMB share is to do it on the Credentials > Local Users GUI. Only one folder can be set to one user at one time so it's a hassle for now.

Is this a fresh install or ported from Core? Basic SMB sharing should work with POSIX1E acltype (though the webui ACL editor has issues).

 

[cyrus104]

There's not really a way to easily "undo" ACL changes. Options are the same as if you had chown -R <name>. The changes are written to disk, and you either have to (1) roll with it or (2) restore from a backup (or ZFS snapshot). The ACL editor for POSIX1E ACLs should be working in the WebUI soon. This portion of the webui is going through a major rewrite to make it overall _much_ more usable.

If you have a snapshot of the dataset from before the migration I do have an ACL restore tool that _might_ work. I wrote it for core and ported to scale, but it hasn't been tested yet.

Thanks for response, right now all of the permissions are basic with one user for SMB and NFS (not the same data sets). Because it's a simple homelab, everything is anonymous user for the most part. I did create a user and group for the dataset permissions that I connect with for my NFS.

 

[LouisB]

This is excellent news...
I will finally be able to move from CORE to SCALE have

 

[Kasazn]

Is this a fresh install or ported from Core? Basic SMB sharing should work with POSIX1E acltype (though the webui ACL editor has issues).

It's a fresh install. :)

 

[Gcon]

Is there any news on what the performance of ZFS will be like on SCALE vs CORE? Reason is I have a 6-disk ZFS array (stipe of mirror'ed vdevs) dedicated as a Nakivo backup box. I was running on the latest TrueNAS and was very happy as virtual machine "screenshot verification" and "flash VM boot" all worked (Nakivo features). Unfortunately Nakivo only have a partial product solution for TrueNAS as they support VM backups on TrueNAS, but not Office365 backups, as it won't let you create a "SaaS" repository. Bummer.
So I rebuilt the box as Ubuntu Linux Server 18.04 (Nakivo don't currently support 20.04), and installed ZFS. Setup the array exactly the same with my stripe of mirror vdevs, ashift=12 for my 8TB 4k/512e drives, and perfomance is woeful. Screenshot verifcations and flash VM boot doesn't work, even with tweaking everything I can with turning off atime, zero compression (software does compression and the LZ4 gets me next to nothing here), and some other tweaks. But ZFS on Ubuntu just doesn't cut it for me for running Nakivo. Granted ZFS is not a supported filesystem for Nakivo on Ubuntu, but I tried anyway.

So now I plan to install SCALE (testing the new beta... non production!), and then put Nakivo in an Ubuntu 18.04 KVM, and mount ZFS storage into Nakivo using NFS for VM backups, and iSCSI for a block-device SaaS repository. If performance isn't great with this, I'll try a similar setup on CORE (with a bhyve VM, plus NFS and iSCSI).

Long story short - how does ZFS performance on SCALE compare to CORE? I'm very wary of ZFS on Linux now after being scarred by my Ubuntu 18.04 experience with poor performance. No bugs but just stuff is noticeably slooooower. THANKS!!

 

[Gcon]

SCALE 21.06-BETA.1 (Angelfish)

2021/06/22

pushed out a week.

 

[LouisB]

take the time you need to make it strong

 

[Kris Moore]

SCALE 21.06-BETA.1 (Angelfish)

2021/06/22

pushed out a week.

Yea, we know that this BETA will be a lot of folks first impression of SCALE, so we're taking that extra week to make it the best it can be :)

 

[Nate W]

Yea, we know that this BETA will be a lot of folks first impression of SCALE, so we're taking that extra week to make it the best it can be :)

Is there a change list published yet? The linked Jira doesn't open due to permissions error :(

 

[Patrick M. Hausen]

Is there a change list published yet? The linked Jira doesn't open due to permissions error :(

You need to log in to JIRA to see the issues.

 

[Kris Moore]

Is there a change list published yet? The linked Jira doesn't open due to permissions error :(

Its a long list in Jira, but yes.

Issue Navigator - iX - Bug Tracker (Jira) (ixsystems.com)

 

[Nate W]

You need to log in to JIRA to see the issues.

I had thought of that. I tried logging in before I posted and same deal.

 

[Nate W]

Its a long list in Jira, but yes.

Issue Navigator - iX - Bug Tracker (Jira) (ixsystems.com)

Perfect, this link worked. Thanks!

 

[usergiven]

I've been reading up on k8s and playing around with it on an ubuntu VM so I can be a little more competent to use things like TrueCharts with Scale. Very much appreciate the work you all have put into it and also appreciate keeping it in the oven as long as it needs to be before beta. Thank you!