Upgrade Central V0.7 to v0.8 'Something went wrong: the server returned an invalid error' and 502 bad gateway

1. What is the problem? Be very detailed.
Yesterday I upgraded ODK central from V0.7 to V0.8 following https://docs.getodk.org/central-upgrade/ everything went well and seemed to install fine. When I went to login to the server I was presented with the error message ‘Something went wrong: the server returned an invalid error’. I then tried to access the server via R and was given a 502 bad gateway error.

2. What app or server are you using and on what device and operating system? Include version numbers.
I’m running ODK central V0.8 on a digital ocean Ubuntu Docker 5:18.09.6 1 vCPUs
1GB / 25GB Disk with only IPV4 enabled. Everything was running fine when V0.7 was running on the same server before the upgrade.

3. What you have you tried to fix the problem?
To try to resolve the issue I have added new users and protoed them to admin to be able to access Collect and still got the ‘Something went wrong: the server returned an invalid error’. I have tried to re-install from the git hub image. I’ve tried removing .server nginx and reinstalling just them.
4. What steps can we take to reproduce the problem?
Run an upgrade from Central V0.7 to V0.8 on Digital Ocean to see if you encounter the same issue.

What, if any changes have you made to the server. Paste the output of the following command.

cd ~/central;
git status;

Also, have you tried a reboot?

1 Like

It'd be great if you could paste in the output that @yanokwa suggests above, as well as trying a reboot. Another thing to look at is the server logs. It'd be useful if you could go through the upgrade instructions again, then type:

docker-compose logs service

(You must still be in the project folder, central, when you run this command.)

Central v0.8 is stricter about invalid forms than v0.7, so if an invalid form was uploaded before the upgrade, there's a possibility that Central would fail to upgrade to v0.8. However, if that's the case, it should be evident in the server logs.

@Matthew_White @yanokwa thank you both for the suggestions. I actually ended up destroying the droplet yesterday and starting again so won’t be able to send the status or service logs. To answer the other question I hadn’t made any additional changes to 0.7 before the upgrade and didn’t receive any warnings about using git stash before the upgrade. I rebooted the droplet a number of times before reporting the issue.