Is it possible to migrate Cloudron from Ubuntu 18.04 to 20.04?
-
@girish
Sorry for the late reply, work is killing me at the moment...All sort of things happened actually after updating on Ubuntu 18.04 from version 6.2.8 to 6.3.4.
A lot of things are covered already in the forums, like the "reboot button" didn't work and so on.
But he main problem i had, was Element's throwing out errors and was offline, on Windows and Android clients. I made a post about it recently.
https://forum.cloudron.io/topic/5291/problem-with-sending-and-receiving-message-s-android-windows/3?_=1626202967742Also i discovered that Nextcloud doesn't sync no more, even not after trying to re-enable sync on the client side, like Nebulon suggested.
Now i'm still running on Ubuntu 18.04 and Cloudron version 6.2.8. and everything seems to be running ok. Except Nextcloud that doesn't sync from the mobile app...
-
@girish
Thanks for the support.
My weekend is started now, so i can finally dig into the problems i had.I'm going to first spin up a new Ubuntu 20.04. VM.
Then install Cloudron version 6.2.8 on that.Restore backup from the old Ubuntu 18.04. version 6.2.8 to the newly created Ubuntu 20.04 VM.
If that's running smooth I'll give it a shot to upgrade to 6.3.4. and see how that goes. I'll report back on that.
If Nextcloud sync is still bugging me after updating, i shall open a separate thread for that one
-
I have done migrating Cloudron to Ubuntu 20.04 and this is what happend.....
As i spinned up the new VM, i setup cloudron to install version 6.2.8. at the command line.
I downloaded the backup config before i turned off the old vm, so i could use the backup config while restoring.This turned out REALLY nice, since my bucket is at DO i could restore from there, after filling in my Id and secret.
After restoring, the Cloudron dashboard quickly showed up online and all the apps were restoring.(edited)
SOLVED!! BECAUSE IT TAKES TIME TO REFRESH THE STATS-->
Ok, so i went for some (basic) checks and discovered that after the restore, the "system info" tab still showed the old values from the old VM. As in disk space was showing the old values.
Strange because i started with a new fresh VM and more disk space assigned to it.
I tried rebooting Cloudron, but that didn't help either. <-- SOLVED!! BECAUSE IT TAKES TIME TO REFRESH THE STATSAlso if i try to update "Collabora Office" (for example), it loses connection (integration) with Nextcloud.
Same is also for "OnlyOffice" that loses connection after updating.
If i restore them to the version before the update, everything is working ok again.Ok so not really smooth running at this point, but this also happend before i did the migration, when it was running on Ubuntu 18.04.
So i thought let's try updating Cloudron to 6.3.4. since i was already on Ubuntu 20.04 now and maybe it could solve my problems from there....Well that also didn't catch the problems i already had, it even became worse.
After the update, Elements was working for like 10/20 minutes and again i could't send message's from the Android app.It seems like everytime i try to update an app or the whole Cloudron environment (version), things are breaking.
I went back to the old Ubuntu 18.04. VM with Cloudron 6.2.8.
That's my only stable Cloudron at the moment.I'm out of idea's at this point and can't detect were it's failing....
Help?
-
@dantheman said in Is it possible to migrate Cloudron from Ubuntu 18.04 to 20.04?:
the "system info" tab still showed the old values from the old VM
I think the stats are not 'live' but are refreshed periodically, so it may take 24hrs for the new VM to report its data.
-
@timconsidine
Now that i'm thinking about that also.
I remember reading something about that in the forums.That's one problem I can skip off my list
-
@dantheman Do you think you can write to us on support@cloudron.io ? I can look into why the update is breaking things.
-
After some investigation, it seems that all the errors are somehow related to OCSP stapling. For some reason, the certs give an error in firefox and android after some time. There is no error in OCSP related errors in nginx logs. Disabling OCSP check in firefox makes things work. Kinda bizzare. Still investigating what the issue is.
I thought this might be time related since NTP was turned off on the server but turning it on using
timedatectl set-ntp ondoesn't help either. -
I created https://git.cloudron.io/cloudron/box/-/issues/796 which will add an option to disable OCSP stapling since we couldn't figure the root cause. The certs itself are fine because I was able to put the certs manually on my server and they seem to work.
