unable to go from core to scale

[hescominsoon]

So this time I went from the latest version of core to the latest version of scale. And it acted nine kinds of weird for a while and then it finally stabilized and I had access to my data for a little bit, then I didn't. It told me the pool was gone and I need upgrade and then it would let me upgrade a non-adesistant pool and then it would disappear and then we'd throw some more errors and then SMB wouldn't work. It threw some errors and iSCSI wouldn't work and finally after about 20 minutes it'd settle down and I apparently have access to everything. We'll see what happens tonight at midnight and see if I have all of my snapshot and replication tasks working. I'm also going to log into my other profile which has all the shares loaded and the iSCSI drive loaded and make sure all my data is actually still there. I do have debugs of before and after and config backups.

 

[hescominsoon]

This rollercoaster makes me VERY wary of taking my exiting core clients to SCALE with existing data.

 

[hescominsoon]

This rollercoaster makes me VERY wary of taking my exiting core clients to SCALE with existing data.

Bug report filed

 

[hescominsoon]

NAS-128064

 

[diskdiddler]

Oh man I was just about to report this..........


and it's not visible, ok fine I'm submitting mine, they can merge. They may be slightly different, mine is very very quickly reproduced with a VM in under 20 minutes.

 

[hescominsoon]

Oh man I was just about to report this..........


and it's not visible, ok fine I'm submitting mine, they can merge. They may be slightly different, mine is very very quickly reproduced with a VM in under 20 minutes.

They closed mine....no resolution offered

 

[diskdiddler]

So what's your current situation? Are you wanting to go back to CORE or switch to SCALE?

I have now reproduced the issue enough to know, I will NEVER test SCALE on my production system at this point, despite reddit people saying otherwise, switching back to CORE seems to break a few small things.

Log in with Atlassian account

Log in to Jira, Confluence, and all other Atlassian Cloud products here. Not an Atlassian user? Sign up for free.

ixsystems.atlassian.net

 

[hescominsoon]

I have my backup server on scale....my primary I will revert to core once my off-site replication completes(which I will start from scratch and restore from scale via ZFS teplication). I have an enterprise on core so will stay with core as my primary until core dies in 2026 ish.

 

[diskdiddler]

Oh wow ok so proper customer not a homelab guy.
Why will you revert, just curious?

I am not convinced CORE will outright die in 18 months, I think I heard of people working towards keeping it alive. Plus, I would assume if it's not internet visible, it may be remain relatively secure with the right policies in place, despite lack of updates.

 

[hescominsoon]

Oh wow ok so proper customer not a homelab guy.
Why will you revert, just curious?

I am not convinced CORE will outright die in 18 months, I think I heard of people working towards keeping it alive. Plus, I would assume if it's not internet visible, it may be remain relatively secure with the right policies in place, despite lack of updates.

No, this is my home lab system which is technically also my corporate production system that I ran this test on. I have not tried this on a critical customer production system. Crazy I may be, stupid I am not. So my backup system's been running scale for years now. I still haven't liked it, and I decided to try their let's migrate from core to scale. And that's where the drama begins . So no I will not be migrating any existing clients to scale from core. Also, BSD 13.3 goes end of life in 2026. and Chris Moore has already announced that there will not be a truenas core based on bsd14. So when BSD 13.3 goes end of life, pretty much means trueness core goes end of life. They may keep it going for a little while longer after that for any customers that still have active support contracts which they said they would do. once that happens there will be no future for truenas core. So it'll be end of life in 2026 and maybe go until 2029 for any contracts that are still active. But there is no long-term future for commercially supported truenas core from IX systems.

 

[Stux]

So what's your current situation? Are you wanting to go back to CORE or switch to SCALE?

I have now reproduced the issue enough to know, I will NEVER test SCALE on my production system at this point, despite reddit people saying otherwise, switching back to CORE seems to break a few small things.

Log in with Atlassian account

Log in to Jira, Confluence, and all other Atlassian Cloud products here. Not an Atlassian user? Sign up for free.

ixsystems.atlassian.net

Which to be fair, is precisely what IX says in the warning...

Which makes it tough/risky to upgrade unfortunately.

When I did one of my ugpraded I mirrored the boot disk first... and then removed one of the disks. Now I have two boot disks... and I upgraded one.

In theory I could switch back to the other... worst case...

Although I'm not sure if Scale changes anything on the pool as part of the upgrade...

A recursive snapshot of the entire pool before the upgrade would probably mean you could rollback everything... altho isn't there a similar feature these days for rolling back major changes?

 

[hescominsoon]

Which to be fair, is precisely what IX says in the warning...

Which makes it tough/risky to upgrade unfortunately.

When I did one of my ugpraded I mirrored the boot disk first... and then removed one of the disks. Now I have two boot disks... and I upgraded one.

In theory I could switch back to the other... worst case...

Although I'm not sure if Scale changes anything on the pool as part of the upgrade...

A recursive snapshot of the entire pool before the upgrade would probably mean you could rollback everything... altho isn't there a similar feature these days for rolling back major changes?

Not one of the times. When I did the core to scale the pool was gone. It was corrupted. There was no recovery in it. The next time that I tried it after restoring everything I was able to go from core to scale the pool disappeared. But then I was able to import the pool and then I had to upgrade it. So I didn't lose anything but it definitely wasn't seamless. So right now I've got a 50% failure rate on absolute total loss of the pool. This is why I have two servers, one backs up the other plus an off-site replication that's the only reason why I tried this is because I knew in a worst case scenario I had two backups, one being off site

 

[ABain]

@hescominsoon, the ticket was closed as duplicate

[NAS-127416] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

This fix went into Dragonfish BETA.1

 

[hescominsoon]

@hescominsoon, the ticket was closed as duplicate

[NAS-127416] - iXsystems TrueNAS Jira

ixsystems.atlassian.net

This fix went into Dragonfish BETA

Roger that.. thanks for the clarification....

 

[ABain]

So what's your current situation? Are you wanting to go back to CORE or switch to SCALE?

I have now reproduced the issue enough to know, I will NEVER test SCALE on my production system at this point, despite reddit people saying otherwise, switching back to CORE seems to break a few small things.

Log in with Atlassian account

Log in to Jira, Confluence, and all other Atlassian Cloud products here. Not an Atlassian user? Sign up for free.

ixsystems.atlassian.net

The migration guide clearly states that it is a one way process going from core to scale, this is not supported:

Preparing to Migrate

Guides FreeBSD-based TrueNAS users through preparation elements and steps before beginning the one-way FreeBSD- to Linux-based TrueNAS migration process.

www.truenas.com

 

[hescominsoon]

A lot of people aren't reading what I posted. I tried to go from core to scale and it destroyed my pools. Not once but twice. The third time I tried it I had to start from absolute scratch. Going from core to scale only worked when I started from absolute ground zero. Even then I had issues with the pool. When I went from core to scale the third time, it came up and said that the pool wasn't there at all. This time it allowed me to import the pool instead of throwing that weird error. Once I imported the pool then they gave me the option to upgrade the pool. Once I did that I had scale up and running. Now I had tried this once going to twice. I'm sorry. Going to Cobia. I then the third time went to dragon fish RC. Dragon fish RC at least allowed me to import the pool manually, but it still didn't convert it on the fly.

 

[ABain]

@hescominsoon agreed, if you have a debug from the dragonfish upgrade please drop it to me so I can get it reviewed.

 

[hescominsoon]

I already dropped multiple debugs at various stages of the process in the ticket that I mentioned above which has since been closed.

I can't seem to reply to your reply dude it's okay. I've got the system up and running on dragon fish RC and I'm just letting it go at this point we'll see what happens when I go from RC to release for right now I would just consider this a warning and if the fix is there then that's great. I'm briefly converting all of my home/office systems to scale since there's a no real future for core unfortunately.

 

[ABain]

I can see a debug from Cobia where the unique constraint issue impacted the upgrade, which was addressed in the BETA.1 of dragonfish, however i am not seeing one from the dragonfish upgrade with the secondary issue you mentioned, if you do have a debug from and would like us to take a look let me know.

I then the third time went to dragon fish RC. Dragon fish RC at least allowed me to import the pool manually, but it still didn't convert it on the fly.