TrueNAS CORE 13.3-RELEASE in August

Announcements
1

TrueNAS 13.0 has proven to be the most reliable and highest-quality platform for traditional primary storage use cases. Community users looking for incremental fixes and changes to their stable storage platform will have the option of moving to TrueNAS 13.3. We plan to deliver TrueNAS CORE 13.3-RELEASE during the week of August 12, 2024.

TrueNAS CORE 13.3 will include the following updates:

  • FreeBSD 13.3
  • OpenZFS 2.2.3
  • Samba v4.19
  • Updates to SMART, Network UPS Tools (NUT), and other services
  • Various security and bug fixes

FreeNAS and TrueNAS CORE were originally developed using FreeBSD as their underlying OS. Three years ago, TrueNAS began its Linux journey by introducing TrueNAS SCALE. This expanded its potential community, broadened and simplified support for the latest hardware, and opened the door to new possibilities for the software. Today, TrueNAS SCALE 24.04 has matched the quality achieved with TrueNAS 13.0-U6 and has more users.

Bringing forward the 13.3 RELEASE date

TrueNAS CORE’s focus continues to be ensuring storage reliability, stability, and security for existing users. Considering its macro lifecycle, TrueNAS CORE is now in a sustaining engineering phase within the TrueNAS project.

There have been two BETA versions of 13.3. Both have relatively few software issues but have not been widely tested. After some consideration, we decided that TrueNAS 13.3 is safe for community use and provides sufficient value, so it should be moved ahead to RELEASE status.

There are only four minor bug fixes at this stage, and the final QA cycle will be completed within two weeks.

Who Should Use TrueNAS 13.3?

Only current TrueNAS CORE 13.0 community users should consider TrueNAS 13.3. TrueNAS CORE 13.0 users will have three choices:

Remain on 13.0: TrueNAS 13.0 is battle-tested and very reliable. It works well if your predominant need is file, block, or object storage and your system doesn’t require any updated feature functionality. Security and bug hotfixes will be provided where necessary for the storage use case.

Upgrade to 13.3: TrueNAS 13.3 includes updates to FreeBSD 13.3, Jails, Bhyve improvements, OpenZFS, and Samba. It addresses some security issues and FreeBSD/VM compatibility issues. Upgrade only if your existing system has issues specifically addressed in the 13.3 release.

Sidegrade to SCALE 24.04.2: TrueNAS SCALE is where new features and updated components are actively developed and tested. TrueNAS on Linux enables us to rapidly deliver a more feature-rich, stable, and easier-to-use storage product for users and customers alike. This includes the ability for TrueNAS to use a much wider variety of hardware, applications, and configurations.

TrueNAS CORE users will always be able to “sidegrade” to SCALE if and when they’re ready. This can be before or after an upgrade to TrueNAS 13.3.

TrueNAS Enterprise will stay on TrueNAS 13.0

TrueNAS 13.3 will not be available for Enterprise system upgrades. Enterprise users need a battle-tested system and don’t leverage the Jails or VM compatibility. Staying on 13.0-U6 is the safer and well-supported path for mission-critical storage needs.

TrueNAS Enterprise customers will always be fully supported for the duration of their support contract regardless of their software version. TrueNAS 13.0 will continue to be supported, and hotfixes will be provided for critical bugs or security issues. Staying on 13.0-U6 is often the easy and wise decision for production systems.

TrueNAS 24.04 already ships by default on Enterprise products, like the TrueNAS F-Series (HA All-NVMe) and H-Series (HA Edge Hybrid). If your organization is considering a sidegrade to the SCALE-based software now or in the future, as many customers already have, please contact iXsystems Support so that we can assist you in the decision-making and upgrade process.

When Should I Migrate?

If you are deploying a new TrueNAS system, we recommend beginning with TrueNAS SCALE. There is more added functionality, vastly broader support for hardware, catalogs of Apps, better performance on most workloads, and an improved Web UI, all of which make managing TrueNAS easier than ever.

TrueNAS 13.0 users looking for the new capabilities outlined above can sidegrade to TrueNAS SCALE anytime, preserving data and essential NAS functionality such as SMB, NFS, iSCSI, and VMs - with the primary exception being Jails.

TrueNAS 24.04 “Dragonfish” includes early support for Sandboxes, which provide jail-like capabilities using systemd nspawn containers. TrueNAS 24.10 “Electric Eel” also provides a native Docker Compose environment that vastly improves running applications with lower overheads and opens the door to more complex setups. Further improvements to Docker applications and Linux Containers are planned for early 2025.

There will be another post when TrueNAS CORE 13.3 RELEASE is available.

For current software recommendations, always review the Software Status page for recommendations based on your profile.

10 Likes
2

Thanks @Captain_Morgan this is very pleasing to hear.

3

I hope this one makes it into the release. It’s the only bug preventing Mac users from staying on 13.0 and it was still broken, albeit differently, on 13.3 BETA2

https://ixsystems.atlassian.net/browse/NAS-130169

On BETA2 I’m also reproducing intermittent issue where amount of wired memory increases unbounded. I was unable to consistently reproduce yet, and therefore haven’t filed a bug. It may or may not be related to collectd.

And lastly, there needs something to be done to incentives users to test early releases. It’s in their best interest to ensure their usecases work. With Boot Environments FreeBSD offers testing beta releases should have been a no-brainer, and yet, here we are.

(Rant: I still feel freeBSD is far superior to Linux in every respect, that Linux won the world of all the wrong reasons, and while I understand why iXSystems had to develop scale— they have business to run — I’m staying with TrueNAS core forever, or until it is no longer supported, at which point I’ll setup FreeBSD myself, enough to support my usecases. I don’t see myself ever tolerating Linux, not until there is no other choice. /rant).

4

I’ll check on NAS-130169. There is a code fix that may been submitted.

I agree that Boot environments do simplify BETA, but we are reluctant to force users to test on production systems. If we had renamed BETA.2 as RC.1, there may have been more testing.

Re Rant: I agree that FreeBSD has advantages on some issues. Jails, ZFS memory, iSCSI stack (iX wrote that). However, there are other areas where Linux is much easier (turnkey applications that survive kernel updates, NFS v4, file change notification etc). As you point out, its the volume of the Linux market and availability of software that is hard to ignore.

The 13.3 release is about sustaining CORE for users like you. Are there any specific features you use in CORE that are not available in SCALE?

3 Likes
5

Mostly a sandbox experience on-par with CORE’s jails. The feeling I get from threads about app deployment in Docker is “extra networking headaches, and file-permissions problems when you have more than one app that accesses the same files”.

6

Not sure if this news to you:

Future plans seem to be to integrate sandbox support into the GUI more, but it’s there right now, and sufficiently powerful for many scenarios.

7

I would argue that’s currently not on par with CORE since it’s not integrated in the WebUI, not is set on stone it will be (SCALE has a history of things not going as planned).

My perception is that the RAM management is still different and with a few caveats, but it might be just a perception. I have not spun up a recent SCALE release for estensive testing yet.

And as written above, I also have the perception of SCALE’s development being a bit flaky… albeit not always for iX’s direct responsability.

My 0,02 €.

8

As long as there is any test that proves that TNS is better than TNC 13.0 U6 in all aspects, I will start migrating to TNS, but I think it is not the case at least now. :melting_face:

9

And until will be possible continue use jails :melting_face:

10

Speaking for myself, jailmaker was a step toward feeling OK about migrating from core if everything else was in order and I felt good about TrueNAS apps. Minimum viable product to mostly remove a roadblock if all of the affirmative reasons to switch were holding up their end of the bargain.

But it’s not on the GUI, the developer is trying to hand it off, and there’s not clarity of what the future holds for Linux containers on Scale.

Indeed everything is in flux with the next release V1 of a new app architecture, jailmaker there but no clarity on what form “official” sandbox containers might take.

So better to let the dust settle so I don’t need to migrate everything twice.

11
12

Which in the end I believe it’s the opinion of most CORE users[1]: we do not yet see SCALE as an alternative to a rock-solid solution.

  1. at least those that do not flatly reject Linux. ↩︎

2 Likes
13

Mostly a sandbox experience on-par with CORE’s jail manager iocage (<-- fixed that for you). Sandboxes are probably–most likely–not on-par with how current FreeBSD jail managers are. Straight up jails on FreeBSD is dead simple and the technology is baked in, vetted, and proven, where on Linux they are an afterthought and not standard (there are even several different methods currently).

Newer jail managers are far more than just simple create, stop, start.

1 Like
14

:point_up_2:

1 Like
15

We’ve been told there wasn’t enough innovation on BSD but that’s just a FUD pitch. If Docker type apps were desired then a simple change to the jail manager could have been brought to CORE or even Docker itself. This would have been arguably easier/better on BSD because of things like single repositories, automatic dependencies and ABI. On SCALE, you’re pushing/trying so hard when you don’t have to and it’s honestly a bit weird. The simple fact is, iX is going after cash. I suspect Linux help is easier to come by. I cannot believe hardware and apps were the only deciding factors to drop BSD because, from where I sit, apps has been a dumpter-fire on SCALE and that tells me iX doesn’t have a solid plan and are just winging it (which I find hard to believe!); This tells me the cash (from Linux) is big and they were willing to jump in feet first (the new ‘HexOS deal’ is an indication here).

1 Like
16

Two Feature Requests about “integrating jailmaker” and “LXC/Incus” were closed with the indication that developers would be considering the best way forward. So there’s a declared interest for going on with sandboxes/system containes/“Linux jails”/whatever-you-call-them, but the future dircetion is unknown.

17

As far as I know, iX is not a no-profit organization.
They told us they changed because their Enterprise customers, the ones who pay, are not interested in BSD and want Linux.
They also told us that the BSD behind CORE was challenging to maintain.

1 Like
18

Agreed. They (iX) needs money. Not a point against them for that at all!

They’ve also said Enterprise customers don’t use jails/apps. So why jump on K3S or Docker or whatever? And I just don’t believe they would have a strong opinion on BSD/Linux if they don’t run apps. -i.e. “BSD is rock solid, yes however Linux could also be if we change; let’s change to Linux (and incur overhead)!” seems like something no business would ever say.

Of course it’s challenging to maintain (what isn’t); you cannot tell me Linux is easier in all aspects to maintain. The changes coming down the pipe from systemd will keep developers on their toes. Kernel and lib changes causing programs to not compile is another big topic (at least BSD packages compile for the most part).

It’s just weird to me. Yes, there are grains of truth in all points but it’s just odd when I step back.

1 Like
19

We have to take some statements at face value.
I can only point out that a lot of work was (and is) being put in for achieving the same features BSD offered while using Linux… the most noticeable is the RAM/ARC management due to the… differencies in memory handling.

Do I see SCALE as a valid alternative to CORE? No.
Do I use it? No.
Do I pay for either? No.
Does my opinion matter to iX? Would be surprised to have a positive answer.
Did iX manage the whole thing badly? Absolutely.
Do I trust what iX says? Mostly so since I am in no position to contest what they affirm.
Will I use SCALE? Perhaps in a few years.

1 Like
20

Have they though?

Because actually they said the opposite

1 Like