Re: Re: [GENERAL] How duplicate data produce when a UNIQUE index exite ?

Greg Smith <greg(at)2ndquadrant(dot)com> writes:
> That GCC string suggests this is a RHEL3 system, which would have
> shipped with PostgreSQL 7.3. Not sure how they got 8.1 onto there, but a
> later 8.1 is certainly a useful first step to take here, before they get
> any more corruption from that ancient version when trying to fix things.
> Compiling PostgreSQL on RHEL3 from source has some fun challenges;
> http://markmail.org/message/2bclakrjfxtgwcge#query:+page:1+mid:jhj3yf7zxfmsi22i+state:results
> covers the main one I'm aware of. And unfortunately
> http://yum.pgsqlrpms.org/8.1/redhat/rhel3.0/ seems to be empty, so a
> source build may be the only good route to get a newer version onto there.

The least painful route for a user of an RPM build would be to grab the
latest SRPM they can find and then modify that specfile to reference the
newer tarball. This should be a relatively mechanical thing (although
you might have to adjust some of the applied patches). The advantage
over just doing a naive build from source is that the update would
install into the right places, play nicely with the package system etc.

regards, tom lane