However, my point is that some Enterprise releases, like RedHat, will not be back porting everything. This means new drivers might not make it to RHEL until you upgrade. Even somethings will get dropped, (like BTRFS was from RHEL7), because of it’s reliance on kernel updates to get bug fixes and new features.
That last is one reason why I think some drivers and file systems should be developed outside of the kernel source tree. It may be a pain for these to support multiple kernels, but it also means that these drivers and file systems actually get fixed across multiple kernels, without relying on someone else to do the back porting. In this case, I mean drivers and file systems under active development, like nVIDIA and OpenZFS.
Enterprises have always operated 6-12 months behind Community releases and generally do upgrade every 1-3 years based on need. They get professional advice and assistance from our Support team. It’s a very different set of requirements and experience from Community.
It seems somewhat ironic that iX have made CORE the closest thing to a LTS by not developing it anymore but still maintaining security fixes for now. That makes it a very attractive proposition for some people.
Does iX know and if so are we allowed to know when CORE with officially go EOL? I mean an actual date in the diary?
If they they are 6-12 months behind, then per definition they would also need to update every 6-12 months. Which seems unlikely and the 1-3 years that you also mention seems more realistic. How do you then reconcile that with ”We are generally backporting at least 1 version” (with reference to security issues)? With 6 month release schedules, towards enterprise you would need to actively maintain >5 versions in reality based on what you said.
Is the reality rather that the majority of ”enterprise”, or shall we call them paying customers, are still on mostly old versions, including Core? And you deal with them on case by case basis bilaterally.
Correction… the new customers start using a release 6 months behind and then update every 1-3 years (typical). You can see this is the “mission critical” row of the Software Status page.
Agreed its like an LTS release. We don’t have an official EOL date. There are still about 35% of users on 13.0.x and a higher % of Enterprise users. We no longer ship new systems with 13.0, but we are committed to the success of the existing customers. We’d also like to make sure CORE users have a reliably good experience when upgrading to Community Edition.
13.0-U6.7 is a solid and very reliable release with large amounts of testing and validation.
So Kris mentioned on the most recent T3 Tech Talk that when a version reaches its .2 release thats when iX officially stop supporting the previous version.
24.10.2 was released in Jan 25
25.10.2 is expected around Jan 26
So I guess if I purchased a TN appliance between let’s say between Feb 25 - Dec 25 it would ship with 24.10.2 (or whatever the latest version of 24.10 is at the time)?
Now if I was ‘lucky’ enough to make my purchase in Feb 25 I would get almost a whole year before I would be forced to upgrade or take the risk of running an unsupported version. But if I was unfortunate enough to purchase in Dec 25 I would get about a month before I was in that situation?
I appreciate no matter what the cadence was you would always have situations like this however the faster the cadence the more common they become.
Historically I believe TrueNAS used to have an ‘Enterprise’ version and that was much more conservative with its updates. Would you ever think about doing that again and dare I say allow the community to use it?
QNAP most certainly isn’t considered to be enterprise anything, at the very least due to their poor security record. However, the hardware’s not bad and runs TrueNAS SCALE very nicely, making that combination a great consumer-level choice.
You are making an invalid assumption. We support Enterprise customers on their release and don’t force them to update to an immature release. We can do private manual or released hot fixes as needed. They are never running an “unsupported version”, but we may recommend an update to a more recent mature version.
Enterprise upgrades are more reliable because they are known hardware configurations and can be pre-tested as an appliance.
This is why I basically don’t update TrueNAS. I once had my system drive go down and when I tried to install the version I had previously - after two days of losing all my hair I discovered they introduced an installation bug that rendered it impossible to install. I then had to roll back with trial and error until I found one that worked (Dragonfish-24.04.1). I will probably get around to testing a new version at some point but I cannot stand the uncertainty - that isn’t the only issue I’ve had it’s just the most memorable.
TrueNAS is the only software I hesitate to update - I’m for sure not the only one - that should tell you something. Meanwhile I’m over here looking forward to my monthly Home Assistant update and adopting immediately.
You are what we would classify as a Conservative user. There is nothing wrong with that. Just thank all the early adopters who find the bugs and help us resolve them.
The only challenge I have with the six-month cadence is the Nvidia drivers which, as of 25.04.2, are still at 550.142 which was released last December. The current version of the drivers is 570.169 which was released on June 17th. Between December and now, the 50xx series of cards was released, and right now, they aren’t supported in TrueNAS SCALE because the drivers are attached to the major releases. A modest 5060 card would still help with transcoding in Plex, but isn’t supported yet. It doesn’t affect me directly because I put in an old 1080 card in my TrueNAS box and a new 5060 in one of my secondary computers to assist with video encoding.
That’s the rub though - iXsystems releases these new experimental features to the user community precisely because we have been willing to be the guinea pigs to see if they actually work. It’s sort of the unwritten contract - we get a updating but free OS / Application, the paying enterprise customers get something that usually works.
And wasn’t that a major driver for moving from FreeBSD to linux, as well? As in it was hard enough to get new drivers for Linux, getting drivers for FreeBSD was pulling hen’s teeth.
Spot on. The release notes for the last couple of releases contained ample warnings for folk who were using Apps or VMs. Basically, while some transitions were attempted, the language made it clear that one could lose it all and to only attempt updates if one had ample time and documentation to recreate that which would likely be lost.
We have all had issues with iXsystems on occasion, such as asking for feedback and blithely ignoring same when it didn’t seem to suit their business strategy, the vibes of the developer reviewing the bug ticket, or other factors.
But that is also their right. Effectively, the only real power we have over iXsystems is to vote with our feet and transition to a different code base. As you have also pointed out, that can be fraught with even more risk. For all the faults and foibles associated with the Apps, VMs, etc. the core underpinnings as a file server are usually solid.
This is the kind of stuff that shouldn’t happen, just like some of the SSL bugs that @dan has pointed out in the past. These are supposedly stable features not operating as expected, which have applications all the way into enterprise-land…
I think this simply comes down to your target audience. Where things get confusing is the lines between Enterprise and Community users.
Personally I can’t complain as I’ve essentially used the Community versions for many years and it’s served me very well indeed and for that I’m very thankful. Having said that if I was an Enterprise customer I would defo want an LTS version.
Not so much a dilemma as a choice to be made, and that needs to be rooted in what kind of product you are building and to what standard. Rolling 3 month release of a NAS / storage management solution in order to incorporate the very latest third party GPU drivers? Even beginning to consider that starts venturing into the absurd IMO. Feedback in this thread and others has rather been pretty clear that you should be considering quality of core product over new features at this point.
Its is “True” that TrueNAS straddles the Enterprise and Community worlds. We do that deliberately and unapologetically.
The Community gets free software from a professional engineering organization. We get a lot of testing, ideas and feedback. Our software matures much faster. The forums help resolve issue (free support).
The Enterprise pays for Appliances which use the same software but with a license for additional features like HA. The hardware is well validated and the software is battle-proven by the Community. We have dedicated and professional Support to simplify the process (24x7). They do get an LTS experience and upgrade every 1-3 years. They don’t see the experimental features until they are fully baked (e.g. Enterprise users don’t yet have VMs).
The revenue we collect from the Enterprise version funds future development work which leads to new software features/versions and another round of Community testing. Its a virtuous cycle.
The other major benefit is that the Community software is the way we find Enterprise customers. We have million dollar customers that used Community edition at home and in the lab. They then wanted Enterprise HA and support for production deployments. (Please send us more of them our way).
The 6 month cycle is there to allow efficient and rapid development. For users that don’t want new features, there is an ability to lag behind. We use the software status page to guide users with soft advice.
We’d like to offer a middle ground between free Community and Enterprise appliances, but don’t have that yet. There is a poor business-case for Community-only software unless it’s written and tested by volunteers. We do accept software contributions as long as there is no negative impact to Enterprise users.
The 6-month works well for this business strategy… we can respond quickly to new requirements, but maintain an LTS experience for Enterprise users.
