Comment 5 for bug 235379

Interesting. I have removed, purged and reinstalled the PostgreSQL packages by means of 'aptitude remove --purge ~npostgresql' countless times last night, but any new installation always resulted in the same behaviour I described above.

However, using 'dpkg -P postgresql-8.3 postgresql-client-8.3' as suggested by Martin seems to have done the trick. Reinstalling after that creates the cluster and starts the server.

I can only speculate why this is happening but I suspect it has something to do with the fact that during installation I also requested that PostgreSQL be installed, along with Postfix and OpenSSH. The DBMS server was not running when I first logged in but apparently was in the state I described in my initial post. I then began removing, reinstalling and reconfiguring but nothing worked so I filed the report. Maybe the locale settings were off at installation time and thus the configuration of the cluster failed.

To verify I just installed a new virtual Ubuntu 8.04 server (using KVM) and did *not* request that PostgreSQL be installed. After logging in and upgrading all necessary packages I just issued the command 'sudo aptitude install postgresql-8.3' and everything worked out fine.

I am not sure any more what exactly the bug could be, but I am sure that the freshly installed server did not work the way it should, seeing as it could not and would not start the PostgreSQL server, even though it was requested at installation time.

Anyway, thanks for the input and help, Leonel and Martin.