After Update to 2.7.1 of the app : The database collation 'C' is not supported by Confluence.

[A Former User]

I got a admin message after login:

Database: The database collation 'C' is not supported by Confluence. We recommend utf-8. On Windows you can also use the Windows-1252 collation for your locale.

can this be ignored? Should i change this setting?

Regards Jens

[girish]

Indeed, the default collation in the postgres addon is 'C'

db5b92d92c20f2448fa4a3388cd74ba0af=> select datname, datcollate from pg_database;
              datname               | datcollate 
------------------------------------+------------
 postgres                           | C
 template1                          | C
 template0                          | C
 db5b92d92c20f2448fa4a3388cd74ba0af | C
(4 rows)

My understanding was that apps can set collation on a per table basis anyway and this is why it hasn't been a problem so far. I can switch the default collation to utf8, i guess. Or maybe we need this information in the manifest. I will check what is the best approach.

[girish]

There's an even better command \l.

db5b92d92c20f2448fa4a3388cd74ba0af=> \l
                                                                                   List of databases
                Name                |                Owner                 | Encoding | Collate | Ctype |                               Access privileges                               
------------------------------------+--------------------------------------+----------+---------+-------+-------------------------------------------------------------------------------
 db5b92d92c20f2448fa4a3388cd74ba0af | user5b92d92c20f2448fa4a3388cd74ba0af | UTF8     | C       | C     | user5b92d92c20f2448fa4a3388cd74ba0af=CTc/user5b92d92c20f2448fa4a3388cd74ba0af
 postgres                           | postgres                             | UTF8     | C       | C     | 
 template0                          | postgres                             | UTF8     | C       | C     | =c/postgres                                                                  +
                                    |                                      |          |         |       | postgres=CTc/postgres
 template1                          | postgres                             | UTF8     | C       | C     | postgres=CTc/postgres
(4 rows)

[girish]

I have changed the collation to C.UTF-8 explicitly for the next release. For the moment, you can ignore the warning.

[A Former User]

OK,
thanks for your Support.

[girish]

@jensbee4 In postgres, one cannot change the collation and ctype of an existing database. The change I have made changes the type for new installations. So, what you have to do after the next release is to make a backup of the confluence app and clone it to a new location (this will create a db with the correct collation) and move the clone back to the original location (after you confirm it's working).

[A Former User]

@girish
Hi, the next release is Confluence Package v2.8.0, App v7.8.0?
I will do the Backup and reinstall at Friday.

Regards Jens

[girish]

@jensbee4 you have to wait for next cloudron release. The change is part of cloudron 5.6.2 (see the announcement category).