infra/tickets

Fork 1

packages app is down: `Error communicating with Solr` #13240

New issue

Closed

opened 2026-03-31 09:27:00 +00:00 by ankursinha · 10 comments

ankursinha commented

2026-03-31 09:27:00 +00:00

Description of request

The packages app seems down again. The main page does load, but trying a search gives:

Error communicating with Solr

Example query: https://packages.fedoraproject.org/search?query=truststore

No deadline, whenever convenient please.

### Description of request The packages app seems down again. The main page does load, but trying a search gives: ``` Error communicating with Solr ``` Example query: https://packages.fedoraproject.org/search?query=truststore No deadline, whenever convenient please.

zlopez added the

day-to-day

points

labels

2026-03-31 12:57:50 +00:00

zlopez self-assigned this

2026-03-31 12:58:04 +00:00

zlopez commented

2026-03-31 13:08:51 +00:00

Owner

I tried to restart fedora-packages-static project and it didn't help. So we would probably need maintainer of the app to look at it.
@kevin Do you know who we can reach to regarding this project?

I tried to restart `fedora-packages-static` project and it didn't help. So we would probably need maintainer of the app to look at it. @kevin Do you know who we can reach to regarding this project?

zlopez commented

2026-03-31 13:39:18 +00:00

Owner

Looking at it more I think it's related to the deployment change as this was merged yesterday infra/ansible#3151, something must be incorrectly set in the new deployment config.
@phsmoura Could you look into that?

Looking at it more I think it's related to the deployment change as this was merged yesterday https://forge.fedoraproject.org/infra/ansible/pulls/3151, something must be incorrectly set in the new deployment config. @phsmoura Could you look into that?

kevin commented

2026-03-31 23:44:36 +00:00

Owner

Huh. I checked it after I deployed @phsmoura 's pr... it was running fine?

It also seems to be running fine for me right now?

That might have been when I was moving it from deploymentconfig to deployment and there was some downtime when it had the wrong volume mounts, which I cleaned up...

Huh. I checked it after I deployed @phsmoura 's pr... it was running fine? It also seems to be running fine for me right now? That might have been when I was moving it from deploymentconfig to deployment and there was some downtime when it had the wrong volume mounts, which I cleaned up...

zlopez commented

2026-04-01 08:29:25 +00:00

Owner

The error is still there, so I don't think that helped.

abompard commented

2026-04-01 08:45:19 +00:00

Member

If I refresh the page multiple times, I get the page sometimes. It is possible that some AI scraper has found its way into this app and is overloading Solr's connection limit?

zlopez commented

2026-04-01 15:33:34 +00:00

Owner

I don't see high load on those pods, so I don't think that's it.

kevin commented

2026-04-01 23:21:48 +00:00

Owner

Yeah, I don't understand. It's basically always worked fine when I have checked.

Perhaps we could add a zabbix check, and see if there's a pattern to when the errors appear?

Yeah, I don't understand. It's basically always worked fine when I have checked. Perhaps we could add a zabbix check, and see if there's a pattern to when the errors appear?

zlopez commented

2026-04-02 11:34:14 +00:00

Owner

I played with it a little today, as I found out that the service needs some changes after DeploymentConfig->Deployment change. It blocked queue processing in the-new-hotness and I noticed similar problems in packages, so I fixed it there as well. Hopefully it will help with the issue.