The Future of TrueNAS and what I means for users like me

[nojohnny101]

Hello everyone,
I've been lurking the forums for a while now (since 2005) and have 2 systems that I built myself (one as a replicate target that serves as an offsite back up for the primary).

My use case for TrueNAS has been fairly straightforward with my primary use cases being mounting shares to be used with Macs (currently using SMB shares) and a few jails, namely plex, tautulli, and qbitorrent. These jails were all built by myself using guides on here (I've avoid plugins because of the many documented brittleness)

I've been following the news on the direction that TrueNAS has been moving. I'm currently on TrueNAS core but am going to have some time in a few weeks to rework my setup. I'd love people's opinions on where I should invest my time. Here is what I've gathered from forum discussion and iXSystems announcements/blogs

- people should migrate from core to scale
- jails should become docker containers in scale

Are these correct assumptions?

Now for the opinion part. I simply want to maintain the functionality of my setup above. Namely multiple services, with their own unique IP address and networking (a great feature of the jail system) with the ability to exchange files between the services (like jails do with mount points) and having shares where I can store my personal files.

My worry is scale is "too much" for my simple use case. I'm more than happy to move in the direction and TrueNAS is going but does it make sense in my simple use case?

Appreciate your thoughts

 

[Patrick M. Hausen]

At the current state SCALE does in no way deliver the stability, security, and ease of use of jails. If CORE was ever discontinued, I would for sure leave the TrueNAS world and go back to "naked" FreeBSD and Ansible. As an added bonus that would get me a current version of bhyve.

Then again nobody ever said CORE was to be discontinued, yet.

 

[winnielinnie]

Now for the opinion part. I simply want to maintain the functionality of my setup above. Namely multiple services, with their own unique IP address and networking (a great feature of the jail system) with the ability to exchange files between the services (like jails do with mount points) and having shares where I can store my personal files.

Not to sound blunt, but why migrate to SCALE? You seem happy with Core.

 

[Patrick M. Hausen]

Not to sound blunt, but why migrate to SCALE? You seem happy with Core.

Then again nobody ever said CORE was to be discontinued, yet.

See?

 

[Juan Manuel Palacios]

At the current state SCALE does in no way deliver the stability, security, and ease of use of jails. If CORE was ever discontinued, I would for sure leave the TrueNAS world and go back to "naked" FreeBSD and Ansible. As an added bonus that would get me a current version of bhyve.

Then again nobody ever said CORE was to be discontinued, yet.

If you do that, and care to publish your configs in some kind of infrastructure-as-code fashion, count me in for the maintenance effort! Over the last many years I've devoted my professional energy to that type of work in multiple different ways, plus I'm passionate about FreeBSD!

 

[danb35]

- people should migrate from core to scale

That is not iX' position as a general proposition; they're saying that both CORE and SCALE are going to continue to be viable products for the foreseeable future. Some of us are skeptical, but that's their position.

- jails should become docker containers in scale

I think it's more accurate to say that plugins should move to apps in scale. Leaving aside the question of whether it ever could have been otherwise, plugins on CORE (and on FreeNAS before it) have never really been stable and reliable, and iX has pretty much given up on them. If you want something like a plugin (i.e., you push a button, maybe fill in a few fields to configure, and it's installed and running), the solution is to use apps on SCALE.

Jails, however, are quite stable, and offer some features that apps on SCALE don't. There's no particular reason to move from software you've installed in a jail to an app on SCALE.

 

[Juan Manuel Palacios]

That is not iX' position as a general proposition; they're saying that both CORE and SCALE are going to continue to be viable products for the foreseeable future. Some of us are skeptical, but that's their position.

I think it's more accurate to say that plugins should move to apps in scale. Leaving aside the question of whether it ever could have been otherwise, plugins on CORE (and on FreeNAS before it) have never really been stable and reliable, and iX has pretty much given up on them. If you want something like a plugin (i.e., you push a button, maybe fill in a few fields to configure, and it's installed and running), the solution is to use apps on SCALE.

Jails, however, are quite stable, and offer some features that apps on SCALE don't. There's no particular reason to move from software you've installed in a jail to an app on SCALE.

Plus, as we all here surely know, CORE plugins are nothing more but jails + pkgs + configs, which can always be replicated reliably, even if at a small scale (pun not intended), with system such as ansible, Bastille, etc.

 

[jgreco]

- people should migrate from core to scale

Nobody has said that unless you're desperate for Gluster distributed FS stuff or maybe Kubernetes (which I feel can be done just fine on CORE with a Linux VM). CORE to SCALE is considered, at best, a sidegrade. Definitely NOT an upgrade, quite possibly a downgrade if you need decent ARC behaviour, iSCSI services, etc. CORE is not seeing significant development because they are focused on SCALE, which is rather behind the curve in some ways. CORE is considered very stable, but some aspects such as the jail ecosystem is pretty tragic. This isn't a problem if you roll your own jails and are comfortable in a FreeBSD based packages or ports environment.

 

[nojohnny101]

Thanks everyone for the thoughtful replies. I guess I read it wrong in this post here in the section "Ongoing Support for Plugins". But I guess they were talking about plugins.

Regardless, seems to be the consensus here; stay on CORE, it's not going anywhere in the near or middle term future, and self-setup jails would not benefit from moving to a VM.

I have always stuck with jails because they are incredibly resource efficient and that is what I need being on slightly older hardware. Spinning up multiple VMs seems like it would hurt the performance of all my services to a noticeable degree.

Thanks everyone, feel free to continue to post thoughts. I think I better know my direction though.

 

[jgreco]

Jails in FreeBSD aren't going anywhere. They've been there since the very late 90's. Additionally, you can advocate for some jail-ish implementation that is being discussed for Linux as well. There's a recent thread about it, check posts from @morganL IIRC, and I strongly encourage this because of the lightweight nature as you noted.

 

[HoneyBadger]

Hi @nojohnny101

You've got it correct - the focus of the blog article linked is limited to the eventual sunset of the FreeBSD "Plugins" solution. Since you've been using jails that you created, you're already ahead of the curve as the first option suggests to

1. Migrate their plugins to personally managed jails and remain on CORE

So feel free to use that free time in a few weeks to polish up your current setup - or experiment with SCALE in a separate context or separate system, if you like.