Reaching 500 Internal Server Error with application startup after previous success

[sahith00]

I had previously requested help for starting up LinkedDataHub, and I managed to successfully run and interact with the application. However, I'm now reaching a HTTP Status 500 - Internal Server Error whenever I run the app. Error messages are pointing to various java files, but I'm not sure where and how to access these and find the cause of the error. Please let me know what I can do to fix the issue, thanks!

Error message:

[namedgraph]

Hi Sahith, sorry to hear that. Can you try executing docker image list and pasting any atomgraph/linkeddatahub-related lines? Most importantly the Image ID.

[namedgraph]

This shows that using the latest tag in docker-compose.yml is bad practice... Because we can't easily track image versions that users have installed. We actually tag each new image build with a source commit hash, but don't actually use those tags :/ I might need to make some changes to that tomorrow.

[sahith00]

Hey, thanks for reaching out! I actually updated my fork with the original code, and 500 error is no longer showing up, so the bad practice causing the 500 error may have been fixed in previous commits to the code. Instead, I'm now reaching a 400 error. I didn't spend much time debugging this, so the fix may be pretty simple, sorry for asking you before taking a stab at it myself!

Here's the error message:

Here's the result from running docker image list with the new code:

Thank you for your help!

[namedgraph]

Since you've updated the code, first you also need to update the Docker images by doing:

docker-compose pull

Re. the WebID profile cannot be loaded -- have you reset the database at some point? By doing anything like:

sudo rm -rf data uploads && docker-compose down -v

In that case your SSL certificates are not loaded to the triplestore. Try this for startup:

docker-compose --force-recreate up

And if it still doesn't work then please copy and paste the Docker container log here. You can get it by doing:

docker-compose logs linkeddatahub

[sahith00]

I updated the Docker images, and I ran docker-compose up, which recreated the linkeddatahub_linkeddatahub_1 container. I wasn't sure what you meant by "loaded in the triplestore", but I have reset the database in the past, so I wanted to try the force recreate command at first, but I wasn't able to run docker-compose --force-recreate up, possibly because the command doesn't exist. My SSL certificates are all still in my LinkedDataHub directory, though.

After this, it still wasn't working, so I decided to get the Docker container logs. However, the container logs for me were extremely long (ended up being around 2600 lines). I created a Github Gist for it, so you should be able to find the logs by looking up sahith00/docker_container_logs.txt in the Github Gist search (I wasn't able to link it here, sorry).

I'm not sure where to go from here, or whether I should try restarting the setup process, so please let me know what next steps you think I should take.

Thanks!

[namedgraph]

Hi Sahith. The logs show:

### Uploading the metadata of the owner agent...

Meaning that a WebID agent based on your SSL certificate should be created.

My suspicion is that you are authenticating in the browser with an older/different version of the SSL certificate than you currently have in in the ssl/owner/keystore.p12. Could that be the case?

Which browser are you using? You need to close the browser in order to get a prompt that allows you to choose a new certificate. If that doesn't happen, you might need to purge your choice that the browser remembers. For example for Firefox, click Delete... for any localhost certs:

[namedgraph]

@sahith00 have you resolve the issue?

[sahith00]

Hi, thanks for reaching out! I deleted the existing certificate and re-imported the new certificate as you said, and I'm able to run the app again! Really appreciate all your help!

In the future, would I need to re-import the certificate whenever I update the source code from the main repo?

Thanks!

[namedgraph]

It shouldn't be necessary.

When you wipe your data folder however, the description for the certificate is gone. So you have to reload the datasets (which docker-compose up --force-recreate should do, but also happens after a new image is pulled) or otherwise you'll get the WebID profile cannot be loaded error.

Another situation where this can happen if you have changed the certificate in ssl/owner/keystore.p12 and loaded the datasets, but haven't changed the certificate in your browser.