@dan Bravo. I agree with everything you said.
1 Like
When I looked at the reports section of Scale before and after migrating to EE, there really wasn’t any difference cpu wise.
And when I say that in certain other forums, Kris tends to say where’s the fun in that, why not upgrade now, etc. I guess he needs to encourage more people to upgrade early for testing. I think IX speaks out of both sides of their mouth.
Say it isn’t so.
1 Like
VMs/Instances are marked as “experimental” because we understood software upgrades would not be smooth for VMs. However, we would like to understand where the complications are and then improve the process and documentation.
1 Like
For me, before I migrated to jailmaker from k3s, I was getting around 12-15% CPU usage on a Ryzen 1800x. After, I dropped to 2-3% average. Temps dropped as well.
Or… there are different audiences.
The official information is on the Software Status page.
1st step is to work out what sort of user you are… we take no responsibility for that.
This is perhaps the exact point that @sfatula is making:
In the rah-rah marketing and in the T3 vlogs iX crow about all the great stuff in the new releases (which is fine), but there are few if any caveats or warnings about not upgrading if X or Y and few if any pointers to the Release Notes and Software Status page.
And when a user is notified of an upgrade being available by TrueNAS there are literally neither warnings or caveats that the upgrade will actually break your system nor are there generalised release-specific warnings nor generalise release-agnostic warnings nor even pointers to read the release notes or software status page.
iX simply pointing to a web page or two (after the event) feels akin to Arthur Dent being told that he should have consulted the local-planning notices warning him of an Intergalactic By-Pass that had been posted on the nearby planet in the Alpha-Centuri system. (Douglas Adams always knew how to hit the nail on the head.)
4 Likes
Nailed it!
It’s not inherently more work to maintain an application in a jail than in helm or in docker compose. @victor does it single-handedly:
(not to mention @dan’s continuous work on Nextcloud)
If a curated app catalog is deemed a desirable feature, there must be capacity for lifecycle management. The underlying technology is irrelevant in that respect.
Kind regards,
Patrick
2 Likes
I think if this is an “experimental” feature, then it should be left in beta, or the old feature should be supported in the release for a while.
After I upgraded to 25.04, my Ubuntu vm was completely unusable. No matter I migrated from the old one or installed a new one from ISO, it would not work. It show something like tihs:
efi stub loaded initrd from linux_efi_initrd_media_guid device path
efi stub measured initrd data into pcr 9
Finally, I can only return to 24.10.
3 Likes
They broke every-single-one of their apps in their march/april 2023 refactor. ALL OF THEM.
Then they screwed all of their users by nuking their repo.
And they were utter pricks on their Discord/subreddit.
Up until April 2023, there were like a dozen? official iX apps. Guessing after the Truecharts refactor, they spun up the community repo. And probably started making plans to distance themselves from Truecharts. Gee, I wonder why.
With Docker, you can unshackle yourself from having to deal with anything other than the project images for the apps you want. You know, the same images that iX official, community and truecharts use.
2 Likes
Yes, I know, and said as much–though I think the date was closer to January. That was the one breaking change driven by them.
This has what to do with the quality of their apps?
Or this?
I’m not really interested in rearguing all the issues with TrueCharts; I was responding only to the nonsensical claim that the poor quality of their apps (not the small number or low quality of iX’ apps) fueled the drive to move TrueNAS to Docker-based apps in 24.10. As to the other issues, hope you enjoyed the rant.
The situation with projects frequently/typically distributing docker images and example compose files means that’s not quite true, even if it’s close in principle.
And of course there are (very popular!) apps that don’t fully or officially support BSD (i.e. immich, jellyfin, homeassistant, etc.). That’s not to say I haven’t run all of them in jails (and homeassistant seems to be most popular in vms now) but in all cases more work was needed than just rolling out a compose, and niggling issues or missing features often remained.
1 Like
Yes - they were human and weren’t perfect. But they had their reasons.
But let’s not forget that iX had encouraged them to create a 3rd party catalogue for TrueNAS and then iX pulled the plug on them overnight when they announced that they were dropping Kubernetes completely when they could have provided some parallel running of Kubernetes and Docker (but chose not to).
It isn’t surprising that the TrueCharts devs were pissed with iX, but I agree that dropping the catalogues overnight was targeting trusting TrueNAS users rather than iX.
1 Like
TC aside, I think the gist of all this is that iX needs to decide where they want to take the product. Core was a fine NAS with some arguably shaky bhyve virtualisation layered on top, jails for the FreeBSD tinkers, and ’applications’ (that were silently abandoned). Now, I fully get that iX wants to expand its offering in order protect and grow its market position. While, at the same time, adjusting to shifting technology landscape and trends (Linux vs FreeBSD, the growth of CSPs, hyperconvergence, clustering technologies, etc). Familiar scenario to any product oriented software company.
But this is also where it gets a bit messy. While flapping a bit on the execution in the last couple of years (docker vs kubernetes, API infrastructure, KVM vs Incus, Coral, etc), I think most of that is symptomatic of some deeper more strategic issues - iX being undecided on more fundamental questions around the product as such. Who are the target customers today vs in 2-5 years? How do iX see them using TrueNAS in congruent with other one-prem virtualisation technologies (e.g. Broadcom). Or does it replace them, for certain types of customers, if so who and why would they? How does TrueNAS align with (and complement) the growing trend of moving storage to cloud? What does it offer instead, in order to make that case compelling? Does it want to be a platform in a partner ecosystem, or own the full experience front-to-back? Etc etc.
3 Likes
Yes, this, exactly. It’s appeared increasingly over the past several years that iX just doesn’t have a plan. Certainly part of that results from external factors–Gluster unexpectedly going away, for example. Part of it results from what I and others have noted up-thread–they feel the need to advertise apps/plugins, but have never (in 15+ years) made them a high enough priority to actually put resources into maintaining them. So they mask that by changing apps ecosystems periodically, which surely takes more work than simply maintaining apps under whatever system is in place.
And more onto the topic of this thread, iX has had the same problem with virtualization, which is now on its fourth iteration. First there was the Virtualbox jail, which most here probably don’t remember–around FreeNAS 9.1/9.2, it was possible (and pretty simple) to create a jail that ran Virtualbox, and then you could run whatever OS(s) you liked there. They took that away in a point release. Then, in TRTMNBN,[1] they introduced bhyve. They rapidly (and rightly) scrapped that release, but bhyve made it into FreeNAS 11 pretty quickly.
Then iX got distracted by the shiny, kicked FreeBSD to the curb, and embraced Linux, naturally requiring a new hypervisor (KVM). And to be fair, we’re still using KVM, but the frontend we’re now using is incompatible with the last one in such a way that existing VMs have to be scrapped and recreated (maybe reusing the existing virtual drives).
So–four different apps ecosystems, four different virtualization platforms. “Does iX have a plan” indeed.
The Release That Must Not Be Named, which was to have been FreeNAS 10. ↩︎
1 Like
And they do the same on Reddit. They say upgrade now, look at this or that feature, etc. but never ever mention the experimental vms, never ever mention the software status page, never ever mention check the release notes even, just upgrade now, rah rah. That’s exactly what Kris was doing. Previously Captain Morgan had said they would do better on social media, the last version. Guess not.
This is nothing more than picking on the weakest users, many of whom know very little about Truenas or Linux. And driving them away IMHO when they encounter issues every release.
I tell people it may be best to wait, Kris chimes in with where’s the fun in that. This is setting them up for failure plain and simple.
2 Likes
I agree. iX folks (esp. Kris) love to crow about how the latest release is breaking all records in the number of people upgrading, but it is exactly this target than makes them be over bullish about pushing users to upgrade and the reason that they don’t warn users in the UI and the reason that they then blame users for not reading the release notes etc.
TBH I thought - and still think - that the iX folks are better than this and can do better than this. They are letting themselves down as much as they are letting users down.
3 Likes
I’m guessing the majority of TrueNAS users out there do not frequent these forums. And for those that do, the release announcement of 25.04 is here. Points out bug fixes, optimisations and a few new features. You’d be forgiven to believe this is a representative summary of upgrade changes [1]. Scarce mentioning of “all your VMs will break, and as part of this release, official status has gone from production grade to experimental”… Which is something for iX to reflect on.
[1] Yes, I understand from reading it carefully that the “release announcement” only compares 25.04.0 with 25.04-RC.1. Which is very spurious given the topic of the post (“TrueNAS 25.04.0 now available!”). Logically, the release announcement should relate to the previous production version which the majority of users will be upgrading from, not beta versions that early adopters have been evaluating.
5 Likes
Pretty sure iX wanted to put as much distance between themselves and TC.by that time.
And like many, TC apps were a big draw for Scale. Until it wasn’t.
Like others, would rather see iX have things working 100% and migration paths before rolling them out. But we’re just unpaid beta testers.