Update to Cobia planification

[Janus0006]

Hi all,

Unfortunately, I didn’t gave a lot of love to my system during last year. I’m curently at the latest version of Bluefin (22.12.4.2) and I would like to plan to update to Cobia. I also read that TrueChart will make major changes in January.

I already lived the change in truechart when they changed many things during last update; where many apps stopped wirking for a lot of people and we had to rebuild almost all of them.

I’m looking to know if you can help me to prepare for this upgrade. May I expect application to stop working ? Do you know if I will have to rebuild my apps again ?
Have you advice ? Guide ? References ?

Since the begining, all my apps are running in a dedicated pool, each apps have their own folder for configuratons files. I also have a dedicated IP for my apps (different from my Scale mgmt IP)

I think this setup helped me during last changes.

Are you expecting many problem for me during migration. Do I have things to know ? read ? Any advices ?

Thank you for your help !

 

[FrostyCat]

Your best bet is to give up Truecharts. The project leadership doesn't seem to care about continuity much and users are mostly lab rats at the moment.

 

[sfatula]

If there are apps Truecharts has that are not otherwise available as apps, learn how to "make your own" via running custom apps, i.e., containers. The standard docker containers can be run in almost all cases. You can experiment with this before you update, see if you can convert a single app to the standard container.

 

[apnea]

FWIW I just did the same upgrade @Janus0006 is asking about. Appliance upgrade to Cobia only took a few minutes, the TrueCharts apps however took an age after reboot (hence me looking in the forum and coming across this message). After about 1.5 hrs all 15 apps were up except jellyfin, which was in some kind of endless restart loop. A further reboot fixed that and now everything runs fine. All in all fairly painless for an upgrade of this complexity.

 

[danb35]

TrueCharts has issued a new major release to their base chart, which likely means you'll need to manually upgrade the apps through the UI rather than having Heavyscript do it. There are changes to Ingress configuration, which mainly mean you'll need to edit the configuration for each of your apps to re-enter the cluster issuer name. You shouldn't need to reinstall anything.

 

[apnea]

I think you're referring to this https://truecharts.org/news/persistence-changes? I believe this is slated for Jan. I just upgraded apps with heavyscript without problems.

 

[danb35]

I believe this is slated for Jan.

I think it's already rolled out. A few days after that post, all my apps had a major chart upgrade available, and all of them required that I reenter the certificateIssuer because the clusterManager settings had moved, which is one of the changes that's discussed in that post.

My cron job for heavyscript does minor updates automatically, but not major ones--those I do manually, because they sometimes (though not always) require changes to be made. So this change meant going through my installed apps, upgrading each one, and (for those for which I'm using Ingress--which is almost all of them) updating the Ingress settings. A little tedious, but not difficult.

 

[apnea]

I see what you mean. I have the same settings in my cronjob. When I ran it manually to do all the major chart upgrades it failed. Looking through the logs I found some other errors which in turn led me to upgrade to Cobia. After that upgrade I ran heavyscript again and it performed all the chart upgrades. I don't have ingress on any apps, so perhaps this is just expected behavior.

 

[Janus0006]

So, as I can read, now, to update, I only need time and patience ?
The system will update itself and after couple of minutes/hours, all my apps should have been migrated/updated to the "new" apps system and work "almost" correct ?

 

[Janus0006]

Hi,

To give you an update, I finally updated to Cobia. And, the upgrade goes .. not to bad. To be honest, I did not updated only the software. I also migrated from old hardware (Dual E5-2620 and old Motherboard) to a more recent i3-12100.
As I had new Motherboard, CPU and RAM, I did couple of test on the new hardware before installing it on my existing server. Restoring my config (with not all storage and hardware) and I saw a couple of issues. Il finaly tested to install a Fresh Cobian and restore my Bluefish config backup. It was 50/50.
In the end, I rebulid my server with the new hardware, install a fresh Cobian and reconfigure everything (shares, apps, etc). The new old setup went very well, we can see the system is most mature than it was the first time. I installed it on Alpha and Angelfish version.
I had to reinstall couple of Apps from scratch, like Plex, Deluge and pihole.
As I already abandonned the idea, I didn't try the virtualisation on Cobian
I like the new apps system, even if it's still not perfect (have to refresh the page ton have a real status of the deployment). The rpject is on the right track.

I hope this answers some questions that some people may have.

Thanks

I will also have to change my hardware details in my Signature.
From:
TrueNAS Scale 22.02.0
Rosewill RSV-L4000 full of:
MB Asus Z9PR-D12 with 2x Xeon E5-2620 with 32Gb DDR3 ECC
Blue Pool: RAIDZ1, 3x WD Blue 3Tb - Red Pool: RAIDZ1, 2 VDEV of 3x WD Red 6TB
OS: 2x Kingston SSD 120Gb (mirror-0) - Plex data: Sandisk II SSD 240Gb
Intel 82574L 2x1Gb - Intel X520-DA2 FC 10Gb - NVidia P600
To:
TrueNAS Scale 23.10.1.3
Rosewill RSV-L4000 full of:
MB ASUS Prime B660M-A AC D4 LGA 1700 with Intel i3-12100 with 32 Gb DDR4 (not ECC)
Red Pool: RAIDZ1, 2 VDEV of 3x WD Red 6TB + 1Tb WD Black SN770 NVMe SSD Cache VDEV + 2x Kingston SSD 240Gb for Log VDEV
Plex data: Sandisk II SSD 240Gb
OS: 1x Kingston SSD 120Gb
Intel X520-DA2 FC 10Gb - No more NVidia P600 as I use the iGPU Intel 730

 

[dAlexis]

BTW, my upgrade from latest Bluefin was flawless. Prepared myself to everything, made additional screens with apps setup, but all, including apps, with customized gitea pod, started OK. Two weeks - works OK.