Hello,
I have been using FreeNAS/TRueNAS for the past five years and have been very happy with it. I am running core 13.0-U6.4 and am using several VMs and Jail. Since the latter are not compatible with Scale, I am holding this cross-grade for the time being. However, I understand that I am limited in creating new jails because FreeBSD 13.0 is EOL. I am considering upgrading to 13.3-U1. Before I do so, I wanted to check whether there are any showstoppers that I should wait for to be addressed. Core 13.0 has had several updates recently, while 13.3 has been slow in the release cycle. Please advise on whether to upgrade to 13.3 or etter stay put with 13.0. Thank you.
The upgrade for me go smoothly.
Afaik the most annoyng problems were the dayli boot pool scrub and the threshold not considered on scrubs task for other pools… But they are plan to be resolved.
Also to mention that the buggy shell has been removed from GUI in that branch release.
Thank you @oxyde . The scrub issue seems to be enough to keep me from this update for now although I do not quite understand what you mean with “threshold not considered”. My scrubs are scheduled on a time basis, every 2 weeks. Do you mean that this interval is not honored?
I do not mind the shell being removed, it’s easy to use terminal instead. I did find the shell at times convenient and did not experience any bugs in my light use. But this missing feature will not prevent me from upgrading.
So waiting patiently for U2, hopefully to come before FreeBSD 13.3 is EOL.
no the “fire trigger” of the task is honored, what is ignored is the threshold day set from one scrub to other on the pool. Pratically, everytime the job start, every pool will be scrubbed. This is the jyra ticket opened on september about.
I have bit the bullet and sidegraded with the EEL release. Yep, is a lot of work if you are running (as me) a lot of jails (VMs should migrate smoothly), but i planned all in time (moving each jail one by one on a Alpine VM, with docker, then after sidegrade just some minor change on volumes were needed). If you have other system running Core (like me, as a backup) i have to mention that i have had encounter problem during SSH connection (Core has a bug and can’t connect to system with user that are not root)
1 Like
Thanks again. Understood now regarding the threshold. I will consider.
I understand that I eventually must cross-grade to scale. As you stated, it’s a bit of work, which I do not have time for right now. I also see Scale still as work in progress, too many changes between releases and still too many issues being reported.
I appreciate the tips on how to make the transition over time. I only have a single system, so the cross-grade will be in place, and there is no other core server to talk to. Why did you use Alpine for this?
Needed a light environment ( at the time i was “limited” by a little i3 
) where could just run docker in a VM, and migrate all service were running through jails. That give me the time to plan each service migration plus no downtime (nor jails and VM were up and working).
Honestly the only jail i give up to migrate just watching the network difference was nextcloud, i admit there i just start from scratch in less time and without the fear to miss something.
Data were served to the VM over NFS, every app installed with a yaml on Portainer… After Scale install i just mount directly volumes and some minor network adjustment
Makes perfect sense, and nothing wrong with a light environment. I have two follow up questions:
- Which edition of Alpine did you use? The Virtual edition?
- My understanding is that Scale does not use bhyve. So how can you reuse the VMs created in Core under Scale? (perhaps it is your last sentence that I did not understand.)
I have used the standard x86_64 version, but in hindsight i could have try the virtual version.
This Is a good question. Personally i remember well that, after migration, the VM was there… But i don’t remember if i tried or not to start It.
The documentation says
VM storage and its basic configuration transfer over during a migration, but you need to double-check the VM configuration and the network interface settings specifically before starting the VM.
So should be perfectly legit assume that everything must work, at least with minor changes… But if you wanna be 100% sure, don’t install Scale on your actual boot device, use another disk so in case of need your can replace fast the boot drive and be back online (instead of reinstall Core).
You also have the option of migrating your jails to a FreeBSD server. I just recently moved all my jails except Plex over to a FreeBSD server and am very happy with the move. You’ll have to manually set up many of the things that are available in the GUI in TrueNAS but it is doable. Exporting the jails in TrueNAS and then importing them in FreeBSD (using iocage) is painless.
2 Likes
@oxyde Ok I will then expect some surprises during the cross-grade. I will transition when I have to and at that point there is no return. So I will deal with anything not working in Scale then.
@Jailer Understood that this is an option. However, I would like to stick with a single server because I am in a home setup and powering two servers 24/7 gets expensive. I do in general like TrueNAS. Once Scale is suitable for my use case, this transition is more an inconvenience than anything else.
Can you explain with more detail what you mean by moving each jail to a Alpine VM with docker?
I think you mean you setup a linux VM on your FreeBSD based Core system, transferred each jail to it’s own VM (via configuring that app in the VM and then migrating config/app files?), and then did the TrueNAS Core to SCALE migration on the host?
Did you write up a list of things that need to be manually setup as you did it?
I’m trying to get myself to stop hesitating on what to do about CORE and the jails that have accrued over a decade that I have time to maintain but not rebuild.
I think the unknowns of what I face in the transition is the barrier as with only one system I feel like I’m standing over very dark water, trying to work up the courage to jump out of the sinking boat. (that hasn’t sunk yet and the band is still playing so I could wait longer surely?)
Yes, installing docker and Portainer to manage easily compose file
The config/app file were served to the VM over NFS, after has been replicated from the jail version to a new dataset
Yes, when EEL was released i sidegraded. Some minor changes still needed but everything keep working well
Unfortunately no I did not. I took notes on a few of the things I had to set up and how to do some things but it’s just that, a few notes. About the only thing I’m lacking right now moving from TrueNAS is notifications. I may set that up some day if I get around to it.