A lot of people don’t upgrade for new features but instead they have been told the railway track ahead is running out and you better jump soon if you want to maintain security.
3 Likes
I understood your point. The thing is, we’ve all seen features be completely reworked, changed, removed, etc in the span of even a year. Over the last 3 years there have been several major changes to the apps system. Dragonfish introduces a laughably untested replacement feature replacing Virtualization entirely. The fact is, if someone is waiting just until the feature they want matures or becomes stable, whatever feature they are waiting for might not exist in a years’ time, so better to just be ready to stay on it forever - hence my originally included option.
Edit:
The point I was trying to make before which you might have missed (it was above when I leaned into the guinea pig thing) is the choices we have been given are all bad. No one is trying to shake responsibility for a choice they made by offering you legitimate criticism in this thread.
There is some good news coming… a different topic that 6 month cycles.
In the meantime keep using 24.10.2.2… its a perfectly fine release and recommended for Conservative users.
1 Like
So either a) stick with 24.10 SCALE for another 6 months, as it remains fully supported until 25.10.2 CE or b) move to 25.04.1 CE now and get a feel for the virtualization feature.
Both can be right depending on how one uses the feature.
For me 25.04 CE was right. I could try an LXC and confirm that yep, ZFS on NVMe still sucks, keep doing dedicated machines with xfs for database work.
But that’s because I’m a home user. SMB and two apps (Plex and Foundry VTT), and that’s it. Virtualization only to, well, see above.
Part of the deal we get with community versions of TrueNAS is to have an understanding of our own needs and choose versions accordingly. I know, that’s a tall ask for many.
On the 6 month cadence: It’s aggressive; I like it. I haven’t seen breaking changes. I believe Dan when he says he has: I likely don’t encounter any because I use so few features of the product. Storage, SMB, 2 apps (one custom). My hardware is from 2018 and not changing. There is zero reason to touch it, unless it breaks.
I think a 6 month cadence straddles the divide well: Users who want to be conservative can be updating once a year. 24.10.2 to 25.10.2, that sort of thing. Users who like seeing new features have the option of following the releases more closely.
Some stuff can’t happen in an appliance. It will never be a “rolling release”. Yes that means NVidia drivers won’t be available the moment new hardware is.
I agree that it’s been a bit of a wild ride. From the outside it looks as if iX was trying for hyperconverged, got no traction with that, and gave up on the idea. First the tech to SCALE went (no more k3s, no ceph to replace glusterfs), then the name changed to CE to reflect that horizontal scale was off the table. I can’t fault them for trying for hyperconverged, and changing tack when they realized it was out of reach.
1 Like
Well I’d just like to thank everyone for their contributions on this thread. I personally have found it very insightful.
I think it’s fair to say if we didn’t already know that TrueNAS is used in a variety of different ways and with that comes a variety of different requirements.
As Steve Jobs famously said “You can please some of the people some of the time”.
Big thanks as always to the TrueNAS team.
1 Like
I think that many of the arguments previously made could be made regardless of what version I (or anyone else) currently use of TrueNAS. Would hate to see those get lost in the smokescreen of “just use version XXX.”
I guess my question would be how many issues have people had when upgrading? If the number is very low, then I do not see every 6 months as an issue. That said, I’m early on in this journey, I think I’ve only updated once, I think I’m still on 24.04 (mostly because I’m worried something will get hosed (thanks microsoft for instilling this paranoia into my brain))
Personally? None. Even the k3s to Docker migration went smoothly.
I think you’ve missed the boat now with app migrations, cut off was June 1st for those. If you use apps, prepare for some extra work since you’re still on 24.04.
It’d be reasonable to plan a move to 24.10.2 so you don’t get “stuck”.
While the software doesn’t keep you from jumping versions, the supported upgrade paths are to the next version. So let’s say you plan and plan and then eventually you want to get to 25.10.2. That’s three upgrades: 24.10.2, fix apps; 25.04.2, VMs gone but don’t act on it; 25.10.2, recreate VMs.
That’s one reason I’ve personally been doing the 6 month cadence: It feels a little easier to do one upgrade at a time, rather than do two or three in a row.
But if I had VMs in addition to apps, I’d be on 24.10.2 right now, wait for 25.10.2, then do two upgrades before recreating those VMs from their zvols.
The app migration cutoff was well advertised. As well as it could be: Forums, YouTube, Reddit - not much more that can be done. Hopefully you don’t use apps, and that’s why you felt comfortable staying on 24.04.
Just listened to TTT and have to roll back some of my comments. I didn’t understand the unofficial EOL policy. I thought it was “we deprecate old-old-stable when stable has reached dot two”, e.g. 25.04.2 means 24.04 no longer receives security updates.
But that’s not it. It’s “we deprecate old-stable when stable has reached dot two”, e.g. 25.04.2 means 24.10.2.2 no longer receives security updates.
That puts “VMs are now experimental” into a different light. It also means CE users that want security patches - and you should want them - will upgrade every 6 months.
There’s a teaser for alleviation of VM suffering, more detail next week.
There’s mention that a paid LTS CE could be a thing if enough people want it. So if you want it, can I suggest: Vote for Add an LTS CE subscription option, and leave a comment as to how much you’d pay.
I am genuinely curious whether “we need more than 6 months support” translates to “and we’ll pay for it”. Engineering time ain’t free.
Which is why I didn’t update though it does take reading the forums to determine that. Why would I have to update to Fangtooth and gain those broken features?
That is indeed the dilemma and it sounds from the above that iX’s offered solution going forward is for you to pay them money.
To be clear they haven’t offered it. They said they are thinking of offering it, and they are interested in understanding appetite for it.
Really entirely up to you. Reading Release Notes is of course optional - the pro move is to do it.
Why did I upgrade to 25.04? It had three features I wanted, which didn’t break anything
- It is named CE and resolves the pain of running something called SCALE which cannot …. scale. As k3s and glusterfs are gone
- It supports LXC, which I wanted to play with. I did, it worked, ZFS on NVMe still sucks as expected. I’ll try again when directio is out.
- It will continue to receive security updates after July/August 2025
Back of the napkin math here - let’s say LTS can be handled with one FTE ( a team of at least three ofc, but one FTE’s worth of time). If it’s less that’s great.
Let’s also say we pay that FTE a decent but not exorbitant amount - this is where it gets into pure fantasy land, I don’t know what ixSystems pays. Let’s say 125k. That’s what I paid my SREs, which was good but not top-tier. And let’s say all-in cost is 250k.
If the subscription costs 100/year then 2,500 subscribers means break-even.
The exact numbers will vary, but this might be directionally correct.
1 Like
I do like the idea but then I’ve be a big believer in TrueNAS support outside of the iX hardware line for a very long time now.
I get the inter-connects between the hardware platform and software and how certain features only work or at least work best when they have been developed together but other companies such as Nextcloud manage to sell their support for just software only.
With iX recently stepping away from HW and focusing solely on TN this could bring about changes such as this and I have no doubt their paying client base would increase massively as would their revenue. Don’t get me wrong I think there is still a place for a solution with both hardware and software specifically integrated but I think the bigger market place is in that other area of buy your own hardware, download our software and pay for support. Perhaps support comes with its own LTS version if you wish or maybe you don’t get a choice who knows.
It may be better for iX/TrueNAS to think more holistically about this (which Im sure they are already doing) instead of just focusing on one thing such as LTS release of which I think should defo play a role.
I guess I’m just not here that often and in general I’d look here before youtube or reddit if I’m having an issue but the fact that I missed the “well advertised” cutoff is largely irrelevant.
In the end, my usage is rather basic, I’ve got like 3 SMB shares, a couple jobs set up to back up to a portable drive and run a Jellyfin server. Looks like I’ll have to make notes about my current Jellyfin setup so that I can attempt to duplicate it with a fresh install. Probably not the end of the world (but it probably won’t go perfectly either). Hopefully the shares and rsync jobs don’t get hosed but I’ll screenshot those as well just in case.
Unfortunately, not everyone lives on the TrueNAS forum, my goal is to spend as little time dealing with the server as possible…I have other things to do.
Virtualization was not “experimental” in Electric Eel in 24.10 but was in 25.04. It was unrelated to whether it was in Enterprise.
It is true that we don’t sell “experimental” features to Enterpise.
“Experimental” means we don’t quite know what we can achieve and where all the dragons are. We may change our mind and this may have some impact on users moving forward. It is a warning.
FYI, we changed our mind and added Electric Eel virtualization back in 25.04.2
4 Likes
Wow, that is AWESOME! Now I for one can upgrade once it comes out. Again the decision to wait for later version has paid off!
Thanks for listening.
1 Like