>
> Hi,
> 	I got around the original issue by manually export
> libdir=/var/lib/pgsql2/lib. However I have stumbled across the next issue
> when running the initialization script.
>
> <stdin>:50: PGRES_FATAL_ERROR select "_slony_example".storeNode_int(1,
> 'Master Node'); select "_slony_example".enableNode_int(1);  - ERROR:
> invalid input syntax for type timestamp: "Fri Jan 21 08:10:14.052379 2005
> SGT"
> CONTEXT:  PL/pgSQL function "enablenode_int" line 32 at SQL statement
>
> I have manually run the above two sql in the _slony_example schema, both
> returning 1. So I'm not sure where the source of the error originates
> from.
> The date/time stamp appears a little peculiar. ("Fri Jan 21
> 08:10:14.052379
> 2005 SGT").
>
> I have attached the master schema, hope this helps.

The problem evidently is that the timezone you're using in your
environment is one that PostgreSQL does not recognize.

We use GMT/UTC exclusively throughout our environments so that it is NEVER
ambiguous what time it is.  Note that GMT/UTC does not vary due to
Daylight Savings Time and such.

You can check what the time zone is, I suppose, by typing in the command
"date", which should list a verbose timestamp.

SGT sounds like Singapore time, which is a bit unusual.

You can switch timezones for a particular process by setting the
environment variable TZ.

Thus, either

export TZ=UTC
or
setenv TZ UTC
(depending on whether or not you consider csh programming to be harmful...)

If you at least change TZ to something that PostgreSQL has in its default
time zone database that ought to suffice.

But "best practices" involve using UTC.  (I should probably throw that
into the FAQ or somewhere nearby the NTP sync discussion...)