RACK UI: button to restart Fuseki

[cuddihyge]

Add to RACK UI the capability to restart Fuseki.
Likely via a button on a new Utility tab.

[weisenje]

Created PR #875. Will merge to master if no objections materialize.

[glguy]

This feature isn't working for me, and I'm getting inconsistent results when I try to reproduce the failure with manual command invocations. Is anyone else experiencing this failing?

I get messages in my docker console output

2023-02-13 16:23:07 fuseki.service: 00:23:06 INFO  Server          :: Apache Jena Fuseki 4.6.1
2023-02-13 16:23:07 fuseki.service: 00:23:07 ERROR Server          :: SPARQLServer: Failed to start server: Failed to bind to 0.0.0.0/0.0.0.0:3030
2023-02-13 16:23:07 fuseki.service: java.io.IOException: Failed to bind to 0.0.0.0/0.0.0.0:3030

and when I connect to the docker instance with bash I can see two fuseki processes running (the old one didn't die).

[glguy]

I can't reproduce this reliably. Sometimes it works and sometimes it doesn't. We might end up having to hack around it like:

• systemctl stop fuseki
• check that it stopped - kill it until this check succeeds
• systemctl start fuseki

[Ptival]

I can confirm that it works for me most of the time, but once I got the same behavior as @glguy , so we should keep an eye on this.

A symptom of the bad behavior is that there end up being two fuseki instances: the "old" one, with a very low PID (usually 110), and new ones seem to get spawned at regular interval and die, with increasing PIDs.

[weisenje]

I had not experienced this problem when I worked on it - the "check for stop" seems like a good idea if that's feasible - if not maybe we should consider whether this feature is worth keeping (@cuddihyge I think we added it as a nice-to-have, not crucial)

[weisenje]

Thank you @glguy for looking at this and pushing #903

I just pulled the latest Docker container (confirmed it got the changes in #903), and the Restart Fuseki button is now failing for me. The RACK UI shows "Exception: Error restarting the triple store". In the shell, I can see systemctl status fuseki: inactive

Perhaps we can look at it together to troubleshoot.

[glguy]

I've pushed a commit that fixes it in my local docker image. When I originally tested this I copied the script manually into my docker but it seems that I didn't save the correct one back into the repository.

[weisenje]

@glguy I pulled the latest and "Restart Fuseki" now works for me.

However, if I load an ingestion package that includes optimization, I get "Exception: Optimize script returned exit code 1". Does this work for you?

[glguy]

It works now. same error in two places. It's rather unreliable to have shell scripts calling each other so the the bug ended up duplicated twice <_<

[weisenje]

@glguy I'm seeing restart and optimize both working most of the time, but not all of the time. I don't see any pattern to the failures.

For restart the user can just click restart again and see it succeed. However if optimize errors there is no way to re-trigger optimize without reloading the entire package.

[glguy]

OK. I think I'll need to add additional logging so that I can understand how things are breaking here before I'll know how to fix the restart process.