TrueNAS Core: Soft EOL
Comments
DannyBee
grahamjperrin
> (This is from december and was covered in various places then)
Yeah, the more recent discussions are far more relevant and informative.
Thanks
JonChesterfield
TrueNAS is the FreeBSD NAS appliance thing. ZFS with easy setup, a GUI etc.
This announcement is that they're abandoning that project, replacing it with a version called "Scale", which if you go looking turns out to be Debian.
Bad news for FreeBSD really. I'm sad to hear it but then I also dropped FreeBSD with ZFS for Debian a couple of years ago so I can't really judge the company for the decision.
Pet_Ant
More appropriately slowly sunsetting it because the alternate flavour is seeing growth and the FreeBSD flavour is becoming non-viable.
> The data is showing us that CORE will become non-viable at some point in the future. Without divulging too much, I can say that SCALE is seeing roughly a 5x growth rate compared to CORE and we don't see any reason for that trend to reverse.
CORE (aka FreeBSD) will still get maintenance releases for the foreseeable future. If you are really bothered about it, crowd fund a few developers to continue the project, I'm sure they won't mind. Otherwise, it seems like SCALE or death for them.
CoolCold
Couple of excerpts I found interesting for myself:
> Not officially yet, but I am giving the soft warning that the data is showing us that CORE will become non-viable at some point in the future. Without divulging too much, I can say that SCALE is seeing roughly a 5x growth rate compared to CORE and we don't see any reason for that trend to reverse. Being realistic about it, if somebody is just starting off with TrueNAS today, I'd highly recommend starting with SCALE since that's where the momentum is and is growing.
And
> Since introducing SCALE our growth numbers have shot through the roof, Linux container users far out-scale Jail users
SpecialistK
A year or two ago I moved my outdated Proxmox server over to TrueNAS Core because I was getting more into the BSDs and liked the idea of "less controversial" ZFS support.
Sadly it did feel clunkier in many ways (my Unifi Jail sometimes fails with a vague error message, CPU temp sensors are inaccurate, SMB share auth is a pain), although stability has been rock solid.
I'll be moving back to Proxmox when I rebuild and upgrade disks. It's a shame that a company that was at the forefront of BSD development has decided to move away from it.
gh02t
Jail-based plugins were always half baked and brittle on Core back to when it was still called FreeNAS. Updates seemed to require complicated manual intervention every time and trying to do it through the UI was a death sentence. Scale is attempting to streamline it with container/Kubernetes-based plugins. It's got some rough edges and growing pains, but in my experience so far it has still been an improvement for me even if I liked the simplicity of jails more.
Proxmox is a better platform if you wanna run a fleet of custom VMs etc, but Scale is doing interesting things if your use case is more "plugin"-like. I use both.
SpecialistK
On Proxmox, as much as possible I would just spin up an Alpine or Debian LXC container and install whichever packages I need, so it is more manual than the Jail-based plugins, but not a huge inconvenience.
While typing that I remembered another annoyance I had with TrueNAS, which is that it doesn't let me pick truly custom MAC addresses for my Jails or VMs, which I like for DHCP reservations. I use Coca-Cola's prefix, because it's funny and I know immediately that it's one of my hosted services.
I recall a blog post or thread somewhere describing the issues TrueNAS devs had with Jail plugins, but can't immediately find it again. If I do, I'll link it.
CoolCold
Why it's the shame? Keep being relevant and satisfying users needs are reasons to be proud, not shamed.
When Microsoft added WSL/WSL2 and collaborated with Canonical for smooth UX, is it a shame as well from your perspective?
ls612
As a neophyte to the server/homelab world, why do people have such strong feelings about FreeBSD vs Linux? To me they seem to accomplish much the same things in much the same ways, although my only experience with a BSD like system is MacOS. Am I missing something crucial here?
SpecialistK
I would say the BSDs and Linux differ in two major ways (to a neophyte, at least):
1) is that BSDs use a monorepo - the kernel and userland are all developed by the same team in the same place, rather than GNU coreutils on top of the Linux kernel, packaged by a number of different distros.
So each of the 4 major BSDs (Free, Net, Dragonfly, and Open) are full operating systems with their own teams and priorities. They share code and history (Dragonfly is a fork of Free, Open is a fork of Net; all are derived from 4.4BSD in the early 90s) but have diverged into their own niches. GhostBSD, TrueNAS, opnSense, and GhostBSD are downstream "distros" of FreeBSD.
2) is the license: whereas Linux and GNU use the copyleft Gnu Public License, the BSDs use a permissive license. This means that BSD code can be used in proprietary software (including but not limited to Sony's Playstation OSes and Windows(!!) [ever wonder why the Windows HOSTS file is in such a weird location...?]) and merged into Linux. But GPL code cannot be added to BSD.
Copyleft vs permissive licensing is a bit of a religious disagreement in the FLOSS world.
ls612
OK at least the idea that it’s mostly sectarian disputes makes sense.
I do wonder though, if most Corps want as much leverage and rights for themselves, why did Linux largely “win” in the enterprise world.
SpecialistK
There's a few theories:
* the legal disputes with AT&T meant that Linux was an earlier Free and Open Source Unix-compatible and gained momentum
* the leadership of the respective BSDs are more conservative or "picky" with their merges, requiring sufficient documentation and cross-architecture support
* the BSDs are 4+ different projects rather than one upstream Linux kernel, making contributions harder
* permissive licenses mean that a company using BSD code does not have to announce or share their changes to customers, making BSD adoption harder to notice (unless you read all of the fine print)
halJordan
It's just two different ways to accomplish it. To pick a charged example purely for effect; this is akin to asking why do Republicans and Democrats have such strong feelings on governing the country? Most people are centrists and most people want similar means to get the same outcome for the country.
debian3
I built a truenas server last December and I was hesitating between core and scale. I went with scale since I have been using debian for a while and never saw any reason to switch away from it. Glad I did.
maybeben
Good ole BSDi, always doing great for FreeBSD.
(This is from december and was covered in various places then)
SCALE was always the main future - that has been totally obvious to all customers at this point. It was also obvious there would be some point they were going to either merge them or maintenance mode Core.
Part of this seems like there are people think that "no new features = EOL", whereas it's usually "you get nothing = EOL".
ixSystems is basically guaranteeing support for Core until the users disappear, which is all someone should expect here.
Unlike most companies, TrueNAS has published roadmaps and the bugs associated with features on those roadmaps. See here: https://www.truenas.com/docs/truenasupgrades/ They are pretty clear about everything, and you can see there are no major version upgrades planned for Core and haven't been for a while - just another stable release.
Overall this just feels like "Some people are upset that they wanted to believe ixSystems was going to develop two completely separate products forever, one which used their preferred OS, despite all available evidence to the contrary".
ixSystems has been clear they will support Core until the userbase isn't worth it anymore. Some people seem to have confuesd support with "innovate and move forward".