[Help]: Won't upgrade, won't import fresh from backup

[thefunkygibbon]

Controller Version

5.15.24.19

Describe Your Issue or Question

When i try to do the upgrade db container, it seems to get most of the way fine, but then fails on something and tries to restore and fails.

_omada_logs.txt

So did an export from working 5.15.24.19 and imported into 6.0 but it also fails

server.log

I'm baffled as to what is the problem since it seems to be a working controller currently.
something SEEMS to be corrupt in the config which is causing it but I cannot work out what.

any pointers/known issues relating to this???

appriciate any pointers as I know this is a "me problem" rather than anything to do with the upgrade procedure.
thanks

Expected Behavior

work

Steps to Reproduce

as above

How You're Launching the Container

as above

Container Logs

as above

MongoDB Logs

No response

Additional Context

No response

[mbentley]

It looks like you're trying to run the MongoDB upgrade multiple times according to the logs. Lines 1-72 are a successful upgrade but then it runs again starting on line 73 so the upgrade fails as one would expect.

I don't think I handle the scenario of checking to make sure that the MongoDB is running the correct version before attempting to run a repair so that is probably something I should handle.

[thefunkygibbon]

yeah i thought it was doing that but i figured i must be mistaken. either way, an export/import also doesn't work so its probably something to do with my config.
after a bunch of faffing around i've gotten a bit further but its not ideal.
for example the exported config file (backup) is about 950kb.
i figured it was something to do with the site name (no idea why it would be the case, but some of the logs implied it might be) so i renamed it to a different name. and did a few other minor things like change/remove the coords for the site location etc.
resulting config export is now 100kb!!!
and it imported!!!!
but
none of the devices were adopted already (thinks it was adopted elsewhere) and wasn't able to adopt them because all the passwords i tried didn't work. also it didn't seem to have any of my wlan groups or ssid's or anything like that in any of the config. so it was all for naught.

somehow whilst doing this i managed to brake it even more and had to restore from a docker persistent storage backup i took (should have been done every night but looks like it hadn't ran for 6 months!). so i'm basically going to start 6 afresh and manually create everything. luckily its just a case of wlans and switch port config and nothing too crazy. should have it done within a couple of hours hopefully.

so thanks, but i don't really think you need to do anything which can help me anymore really. :-(

[mbentley]

Well, I added a check to make sure that someone does not attempt to run the upgrade multiple times (#677) to block that in the future. Sorry that the upgrade wasn't more robust to detect scenarios like this.

[thefunkygibbon]

fyi i didn't "run it multiple times" i ran it and noticed that that it started doing the backup process again at the end and looks like it was stuck in a loop.

[mbentley]

Did the container have a restart policy of always by chance? Just want to make sure that I shouldn't be looking for some edge case in the upgrade script that would explain why the upgrade would run multiple times automatically as there should not be any loops that can be triggered (at least I hope not!)

[thefunkygibbon]

sorry, only just seen this reply. i think it defaulted to "unless stopped" so that might explain things.
either way, I can't really test it anymore as i've started afresh now as the upgrade didn't work.