FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11

[Ericloewe]

Ericloewe submitted a new resource:

FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 - Information on how to deal with the migration from FreeNAS Corral to the supported branch of FreeNAS

As per the recent announcement, FreeNAS Corral is effectively not going to see any further development.

A large number of questions has popped up and I will try to answer them here, to avoid duplication and to save everyone some trouble.


What's going on?
FreeNAS Corral had a rocky launch period. iXsystems has decided to focus on the 9.10 branch and use that as the...

Read more about this resource...

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

Oops, good thing for backups...

This resource had to be recovered from backup after an accidental deletion (thanks, XenForo /s).

If you notice anything out of place, please use the discussion thread to make me aware of that.

Unfortunately, all existing reviews were lost. My apologies.

Read the rest of this update entry...

 

[tulth]

If I change aclmode from discard_chmod to restricted on a smb share without sidegrading to 9.10.x will I have problems with the share if I continue to run FreeNAS Corral?

Background:
I located 3 datasets on my FreeNAS Corral server that have aclmode set to discard_chmod.
I would like to run 9.10.x when I get time and a plan together.
If I go ahead and fix these aclmodes will I have problems on Corral?

 

[wblock]

It probably will interfere with VM operation, specifically stuff trying to access SMB shares. Or would, except Corral might silently change it back.

 

[Ericloewe]

There's no real reason to change the aclmode before you're ready to move to 9.x/11, so it's not something I'd worry about.

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

Corral config backups

Added notice that Corral configs aren't usable in 9.x/11.

Read the rest of this update entry...

 

[vikozo6]

Hello
i have been reading this
https://forums.freenas.org/index.ph...ng-from-freenas-corral-to-freenas-9-10-11.36/
i have only a FreeNAS Corral nightlies installation.
i use Docker / VM!
i am a newbie
what should i do, how to update/upgrade without loosing datas?
What will happen if i change from corral to FreeNAS-11 nightlies?
have a nice day
vinc

 

[jerryjharrison]

I have successfully moved from corral back to 9.10.2-U3, however there is a directory on the root of my zfs pool called ".vm-cache" It has a successive set of empty directories under it called "freebsd-12-zfs", and one under that called "os".

I suspect that this is a left over from my corral testing, and want to remove it. When I execute "rm -R .vm_cache" I get the error "Device busy"

Any ideas on how to clean this up?

 

[jerryjharrison]

Update to this. I used umount -f .vm_cache to clear the error, and then all of the directories were deletable, however I then get the following errors, for each of the directories, appearing in the console:

May 4 06:23:07 corral1 collectd[68307]: statvfs(/mnt/zpool1/.vm_cache/freebsd-12-zfs/os) failed: No such file or directory
May 4 06:23:17 corral1 collectd[68307]: statvfs(/mnt/zpool1/.vm_cache/freebsd-12-zfs/os) failed: No such file or directory

By recreating the directories, the error goes away. If I delete the directories again, the error reappears.

Any assistance is appreciated!

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

Further Corral config clarifications

Added several notes that Corral and 9.x / 11 configs are incompatible, making simple train changes fail.

Read the rest of this update entry...

 

[Amir Yalon]

The section about encrypted pools can be updated based on the resolution of issue 23444, or at least provide a link to it. Thanks!

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

Update: Encrypted pools

The migration process for encrypted Corral pools has been somewhat documented in the following bug report:
https://bugs.freenas.org/issues/23444

A link was added to the relevant section.

Read the rest of this update entry...

 

[Ericloewe]

The section about encrypted pools can be updated based on the resolution of issue 23444, or at least provide a link to it. Thanks!

Thanks for the tip! I added a link for now.

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

FreeNAS 11 RC updates

Made a few minor changes to the wording now that FreeNAS 11 is in RC.

Read the rest of this update entry...

 

[Ericloewe]

Ericloewe updated FAQ: Migrating from FreeNAS Corral to FreeNAS 9.10 / 11 with a new update entry:

Migrations for Docker users

Added a few links of interest to Docker users moving to FreeNAS 11.

Read the rest of this update entry...

 

[Ludensen]

Corral's weird ACL mode-question

As I had a faint memory of accidentally turning the ACL mode on (misreading the option) I wanted to do the test described in https://forums.freenas.org/index.ph...ng-from-freenas-corral-to-freenas-9-10-11.36/
My poor HP N54L NAS - after working for a looong time to implement the change - had to undo the change which took just as long...

Now I have 2 questions:
1:
Having changed to and back from "Corral's weird ACL mode" - do I still have a problem?
2:
executing the command:
zfs set aclmode=oops dataset
Unfortunately in my case the message includes discard_chmod.
BUT executing:
zfs set aclmode=restricted dataset
and afterward testing again with:
zfs set aclmode=oops dataset
still gives me a message including discard_chmod.
Actually executing
zfs set aclmode=oops dataset
on ANY of my datasets gives me a message including discard_chmod...
So my guess is, this command can only be used as an initial test.
Is there a command or way that can show the problem has been remedied? :)

Backstory:
Had a NAS running Corral 10.0.4 witout problems - and planning on leaving it like that until 11-stable (If it ain't broken, don't fix it!). ZFS not upgraded and I had a 9.10 boot environment, I could use when upgrading later.
But then my Kingston USB3 boot-stick became corrupted - "only" 5 errors when scrubbed, but I wasn't able to boot any of the 5 boot environments on the stick.

Fun fact:
To test the Corral ACL mode I have installed 10.0.5 on a fresh USB-pen.
Corral 10.0.5 reports 10.0.4 as an upgrade!???
What fact or message on the forum have I overlooked?

 

[Ericloewe]

still gives me a message including discard_chmod.

That's always going to be there on Corral on pools that support it. It's an indication that it might be in use, not a guarantee.

Is there a command or way that can show the problem has been remedied?

Haven't seen anything, but if you set the aclmode, it's going to be fine (assuming it completed successfully, in other words, that it didn't output anything).

 

[Ericloewe]

Corral 10.0.5 reports 10.0.4 as an upgrade!???

They're the exact same build, except for the installer, which got a fix in .5.

 

[Ludensen]

Thank you @Ericloewe
Now I have actually imported the pool on a 9.10 boot-stick without problems. (Nice with backup of older settings files)

if you set the aclmode, it's going to be fine (assuming it completed successfully, in other words, that it didn't output anything).

No output! So I cross my fingers and hope I will not encounter any surprises down the road!!!
Lately I've been haunted by Murphy and his law :-( Slowing my work down here before BSDCan. My backup coverage is a little thin right now. A couple of hdd's in a fireprof safe must do until I get back and can rebuild everything...

 

[CatSailor]

My FreeNAS box lost power recently (I was running the last release candidate 10 prior to the Corral release) and would not load up. After working on it I determined that now may be a good time to move to RC11. Of course the update failed and I ended up re-installing 11 from scratch.

When I logged in to RC11 it attempted to import my existing ZFS pool and this failed. I currently have a situation where looking at the "Reporting" section it shows my directories, however, i am unable to import the Volumes (in the storage tab) and do not have access to the data. I'm a little freaked out about loss of access to the data any suggestions?

*Edit - Restarted box and the volume was able to be imported. Scary but appears to be working now so all good!