Can we PLEASE warn the "General Public" here about the use of TrueCharts?

[Robert Thomspon]

With the current state of once again crapping the bed, could we PLEASE warn new members about the dependence on TrueCharts and their ecosystem? This latest kerfuffle has literal petabytes of data lost amongst their users because of a lack of communication, a lack of transparency and a lack of the actual back ground workings of these systems by their "devs". and this isnt even close to the first time they've done this to the community (causing massive losses of data through failure of the same as above).

I think it's only fair that people be warned that TrueCharts should not ever be deployed in an environment where safe data retention matters and security are so wontonly disregarded. Yes, they are volunteers, volunteering their time to build these systems and apps and interconnectivity. but this is supposed to be open source, not just a cookie cutter version of Apple where "its magic, dont question the how, our devs dont like that" and the end user can pound sand. Their "support" is entirely contained in discord (right there should be huge red flags... but ok), and entirely dependent on catching someone there who knows the exact answer to your problem

TrueCharts has integrated itself to TrueNAS Scale and TrueNAS Coresimply by following the nomenclature already used. I think people have an expectation that the devs of TrueCharts are as competent as the Devs of TrueNAS Scale/TrueNAS Core. And the honest answer is: Ornias is pretty much the only dev over there who understand the intimate interconnectivity between TrueNAS Scale and TrueCharts. The others are VERY knowledgeable, but it almost seems as if theyve all just had their own shit break so often, theyve been forced to learn how to put bandaids on issues.

i guess </rant>

 

[Patrick M. Hausen]

What's TrueScale?

 

[WI_Hedgehog]

With the current state of once again shitting the bed, could we PLEASE warn new members about the dependence on TrueCharts and their ecosystem? This latest kerfuffle has literal petabytes of data lost amongst their users because of a lack of communication, a lack of transparency and a lack of the actual back ground workings of these systems by their "devs". and this isnt even close to the first time they've done this to the community (causing massive losses of data through failure of the same as above).

I think it's only fair that people be warned that TrueCharts should not ever be deployed in an environment where safe data retention matters and security are so wontonly disregarded. Yes, they are volunteers, volunteering their time to build these systems and apps and interconnectivity. but this is supposed to be open source, not just a cookie cutter version of Apple where "its magic, dont question the how, our devs dont like that" and the end user can pound sand. Their "support" is entirely contained in discord (right there should be huge red flags... but ok), and entirely dependent on catching someone there who knows the exact answer to your problem

TrueCharts has integrated itself to TrueScale and TrueNAS simply by following the nomenclature already used. I think people have an expectation that the devs of TrueCharts are as competent as the Devs of TrueScale/TrueNAS. And the honest answer is: Ornias is pretty much the only dev over there who understand the intimate interconnectivity between TrueScale and TrueCharts. The others are VERY knowledgeable, but it almost seems as if theyve all just had their own shit break so often, theyve been forced to learn how to put bandaids on issues.

i guess </rant>

Isn't Scrutiny a front-end GUI for S.M.A.R.T. (with logging and other self-contained features)? How does that result in petabytes of data loss? That should be "look, but don't touch" software (aside from building historical data for its own use).

I'm not [yet] using TrueCharts so am wondering what you're talking about as you've provided no links and search-engine results didn't indicate massive data loss. (I'm picturing corrupted TrueNAS pools based on the tone of your frantic post.)

 

[danb35]

TrueCharts has integrated itself to TrueScale and TrueNAS simply by following the nomenclature already used.

Your rant (your word, not mine) isn't very credible when you can't even get the name of the product right. But what they've done is what anyone can do: create a repository of charts to deploy a variety of applications on a SCALE server. And they've done so in a far more flexible and featureful way than iX have managed to do, despite the latter having far more resources--and they've automated things I've been asking for for many years (specifically, Traefik/ingress).

I agree with you that they could, and should, have been more clear that the latest change to their master chart (which I assume is "the latest kerfuffle" you refer to) would require removing and reinstalling all your apps--but I don't see any reason this should have resulted in loss of a single byte of data, much less the petabytes you claim. I'm not looking forward to deleting and reinstalling all my apps, but the ones I've migrated so far have been relatively painless.

I also agree in not being a fan of Discord as a support channel--I much prefer a forum to a chat room. Though I'll say that the few times I've used it, I had competent help quickly, so it works (admittedly for a small sample size), even if it isn't in the form I'd prefer.

 

[JoshDW19]

See my note here if you're interested in the new Community App Repository: https://www.truenas.com/community/threads/exciting-news-about-scale-apps.109320/

 

[somewhatdamaged]

Agree with the post. I know the guys are incredibly smart, and they do this stuff for free, but to log in and see that none of your apps work, and without the spare time to go digging around discord (with the world's worst search interface) to find a solution, it's really frustrating. Not all users are developers and we're not all as technical as you, so what you may see as an "easy" fix, is really time consuming or hard for others

 

[WI_Hedgehog]

 

[somewhatdamaged]

I rolled back to the previous SCALE version. I have neither the time nor the mental capacity at the moment to spend trying to "sort" this

 

[danb35]

View attachment 65651

I don't have any interest in changing your, or anyone else's, mind on this; surely there's no harm at all, other than perhaps some wasted electrons, in confining your NAS to NAS duties, and running whatever other apps somewhere else. And I don't expect TrueNAS is going to replace my Proxmox cluster.

What initially had me running jails on FreeNAS, and then CORE, was the direct access to the storage--if I'm running Plex, for example, it makes sense for that software to run on the same hardware as the storage, to minimize network bottlenecks (why send it over the network from the file server to the media server, and from there over the network again to the client?). The same is true of lots of other apps, with Nextcloud being a big example, though I've never used it on TrueNAS (which itself is kind of ironic, given the amount of effort I've put into creating, maintaining, and supporting my script).

What TrueCharts brings to the party--and the biggest reason they got me to migrate to SCALE from CORE--aside from the sheer number of apps, is Traefik/ingress. This is something I asked for seven years ago, and far better integrated than I'd even thought to ask for. iX doesn't do it with the official apps, and I guess it remains to be seen whether any of the apps in the official community repo (how does that work, anyway?) will. And the result of that is that I've put some other things on my NAS that aren't particularly data-intensive.

There's no doubt you're taking the safer course of action, but there are logical reasons beyond simply the cool factor to run other apps on the NAS.

 

[Davvo]

Their "support" is entirely contained in discord

As a non-user, that's the only point I feel at ease agreeing to: discord doesn't look the greatest platform for such tasks.

TrueCharts looks a nice project but I prefer core, jails and feeebsd.

Oh, and boo to Apple. Can't forget about that.

 

[WI_Hedgehog]

@danb35 : Network traffic is a great point, I'm not to needing that [yet] as I don't have a Plex or Jellyfin install, though that networking point alone has made me put things on VM in the past. At some point I'll have to consider TrueCharts, thank you for the reminder. (the rest of your post is quite interesting, it's a good read)

 

[NickF]

What TrueCharts brings to the party--and the biggest reason they got me to migrate to SCALE from CORE--aside from the sheer number of apps, is Traefik/ingress. This is something I asked for seven years ago, and far better integrated than I'd even thought to ask for. iX doesn't do it with the official apps, and I guess it remains to be seen whether any of the apps in the official community repo (how does that work, anyway?) will. And the result of that is that I've put some other things on my NAS that aren't particularly data-intensive.

This is a very interesting point. But it's one that raises many questions as to what the future will hold. While TrueCharts, and it's implementation of Traefik is interesting I think it misses the point. While I have absolutely no idea what the devs at IX are planning for the future, let me present you with this:
The current implementation of Traefik in TrueCharts will paint IX into a corner if they tried to implement it in the official release. Think about the massive ramifications something as innocuous as hostpathvalidation caused. The amount of testing that would need to occur and problems the that would result in such a merger would be insane. And thats just talking about all of the weird things it may break on single node systems. Couple that with the goal for "SCALE" to SCALE OUT to multiple nodes, there may be dozens of design decisions that they may make now that looking back will have been REALLY FRIGGEN STUPID given that they were made for a single node.

The way I see it, TrueCharts is a playground. It was never meant to be enterprise stable. If you want enterprise stable, use Core and host your kubernetes stuff on external servers. That is the true status of the reality we live in right now. @WI_Hedgehog is right. If all you care about is stability, use your NAS as a NAS. If you want to play around with free and open source hyper converged software, you have to understand that there are going to be growing pains. I've said this before, and I will say it 1000 times more, my only criticism of SCALE is that I truly think it should be labeled as a BETA or as a Technology Preview, because it really is changing and growing so fast, and the way things are done now may be different as it grows up into the future.

To OP's point, I'm not really sure what else IX should have done. They literally make you manually add the repos for a reason. TrueCharts is not an officially supported entity. The only gripe I have with that is that they probably should have had community charts available as a concept sooner, so as to avoid all of this drama. But hindsight is 20/20.

 

[sfatula]

Moved away from Truecharts for the reasons noted and more. Just using my own apps now. It's no issue for me to not use their Traefik/Ingress. I use Emby just fine. I just assign it it's own IP, no need for a VM. That way the standard ports work.

Have had zero issue with stability and I find Scale quite stable. As stable as anything else, yes, there are always bugs.

 

[WI_Hedgehog]

I'm perhaps abnormally risk-adverse because I come from a Windows Server environment where the motto is, "If it ain't broke, it will be."

That and, "Complexity killed that cat--#1010b times!"

 

[sfatula]

I get that. However, that's one of the reasons I went with running my own containers (extended the official docker apps from whoever wrote any particular software), pretty simple to me. Because, it won't change unless I change it, maybe yearly if some reason to desire the latest version. The developers can't make breaking changes (Truecharts or IX), it's under my control. Since it's under my control, I might never change to a later version of some app. And it likely won't matter. Overall, I spend far less time now than I did with truecharts apps, I spend pretty much zero. For me, so much happier and less things breaking.

An an example, to get some apps working, there is so much extra work to do if you rely on either IX or truecharts. I already had a nextcloud. I used mariadb as I've been doing so for 15 years (well, mysql back then). I didn't want to convert to postgresql. So, avoided that hassle. Then I wanted SSL, well, another hassle with the apps. etc. etc. Emby was even worse to try and make their versions work.

 

[NickF]

I'm perhaps abnormally risk-adverse because I come from a Windows Server environment where the motto is, "If it ain't broke, it will be."

That and, "Complexity killed that cat--#1010b times!"

Nah, I don't think that's even true. I just think that Gen Z is here now and they want all of the features without understanding the risk :)

Because, it won't change unless I change it, maybe yearly if some reason to desire the latest version. The developers can't make breaking changes (Truecharts or IX), it's under my control. Since it's under my control, I might never change to a later version of some app.

While that is theoretically true, if people like OP pressure IX into making insane changes that break how your configuration works, it doesn't matter that you were in control after you apply the latest OS update. ;) Again, see the hostpathvalidation situation.

 

[Trevor68]

It was pretty easy to see this coming, I'm a windows guy and completely linux illiterate, and until this "kafuffle" was docker illiterate too.
I now have my 15 or so truecharts apps happily running on windows 11 VM. I moved everything off slowly over a month while I self taught myself docker compose. So yes, while truecharts currently is fairly broked, to be fair, there was a LOT of warning.

Also I now realise my setup is better for my needs, as I understand how it all works, and can add bits and pieces myself such as watchtower, cron etc etc.

Once the dust settles I will happily use truecharts again, if I can figure out a way around their dependence on forwarding ports.

 

[Patrick M. Hausen]

If you want enterprise stable, use Core and host your kubernetes stuff on external servers.

Or host your standard open source applications in jails and automate with Ansible.

I have mixed feelings about the community repo. Without a clear policy for contributions and a lot of curation by iX or simply "someone" there's a high risk of it going the way of CORE plugins. If you offer application X as a plugin/app/whatever and you cannot guarantee that it's updated in a timely fashion after every single upstream update, users will get disappointed.

So for me it's jails, because the FreeBSD ports/package system does exactly that.

P.S. And of course you will have to ensure updates work smoothly every single time. That looks like the main reason why the plugin system lost traction. "I upgraded my Nextcloud and now it's broken!!!" Well ... applications with a clear indication in their UI that a new version is available are a blessing and a curse.

 

[MisterE2002]

I agree with you that they could, and should, have been more clear that the latest change to their master chart (which I assume is "the latest kerfuffle" you refer to) would require removing and reinstalling all your apps--but I don't see any reason this should have resulted in loss of a single byte of data, much less the petabytes you claim. I'm not looking forward to deleting and reinstalling all my apps, but the ones I've migrated so far have been relatively painless.

Not upgraded yet to latest version. But if they *force* you to uninstall apps and reinstall them then you also delete related PVC data.

 

[danb35]

But if they *force* you to uninstall apps and reinstall them then you also delete related PVC data.

So stop the app, mount the PVC, and back it up (tar -cvjSf /mnt/somewhere/app.tar.bz *). Once the app is reinstalled, reverse the procedure. Kind of tedious, but it's worked for me so far.