0

We are migrating from SonarQube (SQ) 4.5.4 to the latest 6.7 LTS. So far we tried out our upgrades with 6.7 and everything seemed to work fine, now we tried 6.7.1 and the web start up fails with the message

ERROR web[][o.s.s.p.Platform] Web server startup failed: Oracle NLS_CHARACTERSET does not support UTF8: WE8ISO8859P1

It is very clear from the log that we are using WE8ISO8859P1 and SQ is expecting UTF-8. In our infrastructure SQ is living as one of the tenants in single schema, it is impossible for us to migrate to UTF-8; also UTF-8 will be heavy on storage in comparison.

Now my questions are,

  • Why SQ 6.7 was ok with WE8ISO8859P1 character set and SQ 6.7.1 is adamant about UTF-8
  • What can we do now?
  • What about the future releases?
  • Is SonarSource going to force us to migrate?
G. Ann - SonarSource Team
  • 22,346
  • 4
  • 40
  • 76
Shiva
  • 717
  • 9
  • 22

1 Answers1

0

Thanks Jeroen Heier for the answer, however on further digging I found your answer is only half correct. I want to elaborate on why I faced the situation and how I got over it.

SonarQube 5.6 requires UTF-8 as characterset not just SQ 6+

Why SQ 6.7 was ok with WE8ISO8859P1 character set and SQ 6.7.1 is adamant about UTF-8

SonarQube does not check for characterset in existing instance, it is only worried about new installations until Sonar Qube version 6.7.

What can we do now?

If you are more adamant than SonarQube requirement, install SonarQube 4.5.4 first and then point the same database and upgrade. Till now you can only use that till 6.7

What about the future releases? As per this comment, no existing installation should be forced to move to UTF-8. As they have themselves acknowledged that it is painful to migrate to UTF-8 from non-utf.

Is SonarSource going to force us to migrate? Let us hope they don't.

Reference : https://jira.sonarsource.com/browse/SONAR-6171

And I created a post in Google groups : https://groups.google.com/d/msg/sonarqube/jYRzfL5IaBk/8dnPIEnOBAAJ

Shiva
  • 717
  • 9
  • 22