Issues panel not working after upgrade

SonarQube
1

Hello,

Using the community version of Sonarqube, we have upgraded from Sonarqube v6.7.7 to v8.9.8. Our upgrade path was:
v6.7.7 => v7.9.6 => v8.9.8
Open JDK8 => 11 => 11
Potsgres 9.5 => 9.5 => 13

We didn’t get any errors at v7.9.6. But after the upgrade to 8.9.8, the following error message show up when we load the “Issues” screen:
“Results are temporarily unavailable. Indexing of issues is in progress.”

We’re pretty sure the indexation is finished as there are no more background tasks in pending.
We’ve tried to delete the data/es7 to restart the indexation, but we still have the issue.

The exact same upgrade has been done on a second Sonarqube server. We had the same error pop-up on the screen “Issues”, but the re-indexation did work for this one (the db is a lot smaller if that makes any difference)

The “Projects” screen works and we can browse into all the projects to see related issues. Only the “Issues” screen doesn’t work.

Any advices ?

Thanks,
Scott

2

Hey there.

  • I would suggest making sure you’ve run a VACCUM FULL ANALYZE after your Postgres upgrade.

https://community.sonarsource.com/t/sonarqube-becomes-slow-after-postgresql-upgrade/37925/2

  • Did you recycle the entire SonarQube instance (from the command line) after deleting the folder? By this I mean a ./sonar.sh restart or your OS equivalent, as opposed to restarting from the UI.
3

Hi,

Thanks for your help.

  • I would suggest making sure you’ve run a VACCUM FULL ANALYZE after your Postgres upgrade.

Sonar isn’t particularly slow. We can browse the project or the admin panel smoothly. It’s just throwing us the indexation error when trying to browse all the issues. Is it worth a shot anyway ?

  • Did you recycle the entire SonarQube instance (from the command line) after deleting the folder? By this I mean a ./sonar.sh restart or your OS equivalent, as opposed to restarting from the UI.

We restarted the service using command line ( and we only do it this way).

4

I think so.

I would also suggest checking the logs (all log files) after a reindexing to see if there’s any failures or errors, that might help track down the issue.

5

Hi,

It’s fixed, thanks !

VACUUM FULL ANALYZE + re-indexation did the job.

We executed the following command, as postgres user, to execute it on all the tables:

for TABLE in $(psql -d sonar -c "\dt" | grep table | awk -F "|" '{print $2}' | tr -d " "); do psql -d sonar -c "VACUUM FULL ANALYZE $TABLE" ; done

1 Like