TrueNAS history with FreeBSD

It says that TrueNAS Core which is based on FreeBSD will be in maintenance mode based on FreeBSD's releases, and will be so for years. It is also set to match FreeBSD's version numbers. It admitted that the FreeBSD version ran better, faster, ran smoother and needed less systems resources. The claim was that users could upgrade from TrueNAS Core to the Debian version without issues at any time.

It's saying, FreeBSD lacks the ability for features and functionality, unless someone in the community chooses to pick up on them. Also, iXsystems workers spent most of their time on improving FreeBSD rather than innovating. Unfortunately, it described FreeBSD as not picking up and not prioritized.
if people wanted to push FreeBSD forward for the last 15 years, they would have.

FreeNAS started in 2005, and stewardship was handed over to iXsystems in 2009. FreeNAS merges into TrueNas in March 2020. Since July 2020, TrueNAS has had a Linux version. https://www.truenas.com/freenas/

Screenshot from Februrary 2020 of FreeNAS https://web.archive.org/web/20200303161400/https://www.freenas.org/. FreeNAS based on FreeBSD had clients of: Salvation Army, Reuters, US governmental and military branches, United Nations, Disney and a few Universities.

Related: Thread ixs-plans-for-truenas-core-13-3.92883.

Edit: I can see lack of drivers for newer highspeed network hardware and feature functionality of that hardware as a reason to want to do that. Drivers can be imported to FreeBSD. The only problem would be when GPL inclusion would make the rest of the BSD licensed software to go under GPL, by including it. Still, that would be a FreeBSD based system.
 
The Register alludes that iXsystems wants to offer Kubernetes

Although The Reg FOSS desk remains skeptical about whether most of its users actually need Kubernetes, it remains popular and adoption continues to grow. Many companies have hitched their future to this bandwagon and got rich. For iXsystems, it seems to us like a small fish choosing to jump into a big pond, and we are concerned what might happen if the wave of microservices hype crests and dissipates.


So, why does FreeBSD not offer Kubernetes? Because it is not Linux.

But maybe one could hide this with a clever bhyve Linux VM.
 
Sure, let them twiddle with Linux for a bit. I wish them luck with the product. Much of their efforts might be lost in the noise of the Linux ecosystem though.

Its a bit murky but Klara is the commercial spinoff from iXsystems with more of the FreeBSD focus these days? Possibly not too much will change.

So, why does FreeBSD not offer Kubernetes? Because it is not Linux.

But maybe one could hide this with a clever bhyve Linux VM.

So very true. The very definition of Kubernetes (and Docker) is "AMD64 Linux OS".

That's pretty much what Windows (Hyper-V based WSL2) and macOS (xhyve) do anyway.
 
The Register alludes that iXsystems wants to offer Kubernetes

So, why does FreeBSD not offer Kubernetes? Because it is not Linux.

But maybe one could hide this with a clever bhyve Linux VM.
So very true. The very definition of Kubernetes (and Docker) is "AMD64 Linux OS".
Kubernetes supports more container runtimes than just Docker, supports more architectures than just amd64, and for worker nodes supports more Operating Systems than just Linux...Sure it may be that the majority of installations are Linux based on amd64, but I've seen Linux clusters running on Arm and also clusters with Windows worker nodes for Windows containers.
I'm not going to comment on the merits of doing the above, just wanted to point out that should someone want to run FreeBSD containers on Kubernetes it's not impossible but would need some work - and I think Samual Karp's work on runj would probably mean there is less work needed than a few years ago.

My understanding is that the draw to using Linux and Kubernetes for TrueNAS is that it can leverage the huge number of Linux based containers available along with tooling that makes orchestrating and managing those container somewhat easy to manage in programmatic way.
Using FreeBSD + Jails meant that they had to manage the creation and distribution of the application containers themselves often with little to no support from upstream, or use Linux in bhyve which likely caused additional overheads/pains.

There was some talk on a recent episode of 2.5 admins where they discussed this and it was suggested that they will probably keep maintaining the FreeBSD version for as long as they have commercial contracts which are also tied directly to the hardware sold by iXsystems. Once that hardware is EOL and contracts are up, there is no commercial sense to maintain the FreeBSD version and they can kill it.
 
Sure it may be that the majority of installations are Linux based on amd64
Basically this. What are we talking about when it comes to majority? 99.9999% or 99.99999%?

but I've seen Linux clusters running on Arm and also clusters with Windows worker nodes for Windows containers.
Since docker can't be used for non-Linux OSes, what runtime was it using out of interest?

Where did this notion come from?
The website. Specifically products and services.
 
Since docker can't be used for non-Linux OSes, what runtime was it using out of interest?
I don't have access to the cluster and am not very interested in Windows to have cared much about it, but looking at the Kubernetes docs for Windows it seems that it would either have been ContainerD or Marantis Container Runtime (which appears to be an evolution of Docker Engine from what I can see).
 
I don't have access to the cluster and am not very interested in Windows to have cared much about it, but looking at the Kubernetes docs for Windows it seems that it would either have been ContainerD or Marantis Container Runtime (which appears to be an evolution of Docker Engine from what I can see).
Same.

Though if you drill down through all the layers of nonsense; both seem to basically be Docker "for Windows". Which means that "Windows Containers" are just a tiny amd64 Linux sat ontop of Hyper-V controlling minimal Windows VMs.
 
Maybe not clear to casual readers, "you've betrayed us" (pink annotation in the attached screenshot) is, like, an accusation that was directed at iXsystems.

Postscript, for clarity: to my mind, the accusation was unjust.
 

Attachments

  • 1712779993489.png
    1712779993489.png
    62.7 KB · Views: 48
… commercial contracts which are also tied directly to the hardware sold by iXsystems. Once that hardware is EOL and contracts are up, there is no commercial sense to maintain the FreeBSD version and they can kill it.

The optimist in me foresees possible interest in FreeBSD 15.1.
 
The website. Specifically products and services.

Please: any specific URLs?

Will wsoteros (Open Source Community Manager) please, with reference to what's quoted by allanjude@, can you think of an origin for the misunderstanding?



<https://www.google.com/search?q="products+and+services"+site:www.truenas.com+-"community"> finds only one page, relating to Veeam’s products and services.

<https://www.google.com/search?q=Klara+"products+and+services"+site:www.ixsystems.com>:
Appended to the paragraph about longevity and high write tasks:

With TrueNAS CORE scheduled to receive continued updates for stability and security, choosing a quality, reliable boot device is the first step towards a well-built TrueNAS system.

PS for what's pictured below:
 

Attachments

  • 1712819740843.png
    1712819740843.png
    39.7 KB · Views: 22
Seconding what Warner asked, where did you get this idea?
Klara provides commercial support and development services for FreeBSD and ZFS, but, Klara is in no way related to iXsystems.
That is actually great to hear. Multiple separate entities are almost always stronger.

So where I got this idea from (as I mentioned, murky) is actually compounded by a few things on the web:

1) Most outreach events are done together:
It seems like your collaborative work is more commonly advertised than individual work! (granted Fast Dedup was quite a large project).

2) iXsystems specifically advertising some of Klara's articles:
3) Klara's page specifically mentioning TrueNAS on some of their articles (these are a great read btw) i.e:
4) Some of the quotes on this article from iXsystems really makes it sounds like the companies are related:

“Fast Dedup has been a major user desire for many years and we are pleased to see that Klara and iXsystems have developed the feature in-line with our general requirements. We look forward to testing and integrating the new software.”

You stated they are separate entities and since the FreeBSD world is relatively small I can certainly see why collaboration between the two is strong so that all makes sense.

However in the past I have been part of a corporate spin-off and amusingly the techniques used to bootstrap the new company and give it visibility within the sector were extremely similar to what I have seen here.
I knew iXsystems has been around for a while since the BSDi days, so that is what lead me to believe it was Klara which was the spin-off. Turns out not to be :)
 
… in a nutshell, I try to turn on EVERYTHING in a port when I compile it. Packages usually come with a pretty conservative set of options turned on. For example, I turn on all of the audio options available (SNDIO, Pipewire, PulseAudio, JACK, and ALSA...), because they compile no problem, and it's a way to avoid weird lack of sound later on. I also turn on LTO, and turn off the EXAMPLES options.

💯 👍 to people who are willing to take so thorough an approach.

Whilst the given example (audio) is not relevant to things such as TrueNAS CORE, it's probably fair to say that increased testing of non-default options, for an identifiable set of ports, should be a good thing for the FreeBSD Project.

Consider Kris Moore's recent comment in TrueNAS Community Forums …
 
Back
Top