"PostgreSQL Bugs List" <pgsql-bugs (AT) postgresql (DOT) org> writes:
Tom Lane:
Your trace shows no sign of having stopped the postmaster before
re-initdb'ing or starting it again afterwards. I think you've got some
unholy combination of old and new tables inside the postmaster. It'd be
a good idea to try again with those steps included. (I'm not sure just
how Windows handles deletion of directories that are still in use, but
on Unix a pg_ctl stop would no longer suffice to get rid of the old
postmaster, because after the rm -rf you can no longer get to the
directory in which the old postmaster is running. You might be best
advised to reboot.)

Thank you. That did it. May I suggest adding this:


--- INSTALL.old 2004-01-17 13:34:22.000000000 +0200
+++ INSTALL 2004-01-17 13:33:32.000000000 +0200
@@ -15,6 +15,7 @@
mkdir /usr/local/pgsql/data
chown postgres /usr/local/pgsql/data
su - postgres
+ <kill current postmaster. This clears groups and users>
LC_ALL=C /usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data
/usr/local/pgsql/bin/postmaster -D /usr/local/pgsql/data >logfile 2>&1 &
/usr/local/pgsql/bin/createdb test


---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo (AT) postgresql (DOT) org so that your
message can get through to the mailing list cleanly