Hi All,

Due to some historical idiosyncracies in our environment, we have a custom =
8.3.7 database installation built from source. We'd like to install dblink=
into this, however there are some problems with doing so:

1) the 8.3.7 database was built on a CentOS 4 build box that has since gone=
away
2) currently we have only 8.3.9 code built against CentOS 5
3) the GCC compiler on CentOS 4 was quite old
4) possible API changes in dblink between those versions

My question, how risky would it be to copy the dblink.so and .sql files fro=
m the CentOS 5 compilation of Postgres 8.3.9 over to the CentOS 4 compilati=
on of Postgres 8.3.7? If runtime errors result, how severe would they be? =
I.e., would they take down a postgres backend or possibly the postmaster d=
aemon?

Thanks,
David
--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

David Jantzen <djantzen [at] ql2.com> writes:
> Due to some historical idiosyncracies in our environment, we have a custom 8.3.7 database installation built from source. We'd like to install dblink into this, however there are some problems with doing so:

> 1) the 8.3.7 database was built on a CentOS 4 build box that has since gone away
> 2) currently we have only 8.3.9 code built against CentOS 5
> 3) the GCC compiler on CentOS 4 was quite old
> 4) possible API changes in dblink between those versions

> My question, how risky would it be to copy the dblink.so and .sql files from the CentOS 5 compilation of Postgres 8.3.9 over to the CentOS 4 compilation of Postgres 8.3.7? If runtime errors result, how severe would they be? I.e., would they take down a postgres backend or possibly the postmaster daemon?

I think you've got a fundamental problem that you'd better fix. If you
are unable to rebuild the database from source then you are unable to
update --- and you are already three minor versions behind and missing
multiple security and crash-risk bug fixes. You're living on borrowed
time, and NEED to reinstantiate your ability to build for that platform.
Or move to a newer one.

FWIW, you could probably get back a RHEL4 build environment pretty
cheaply by setting up a suitable "mock" chroot on a recent Fedora
version (installed on the same type of hardware).

As for your direct question: no, I wouldn't count on that to work.
RHEL4 and RHEL5 had different glibc versions didn't they?

regards, tom lane

--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

Thanks for the detailed response. We'll look at upgrading to 8.3.9 from an=
RPM.

On Mar 15, 2010, at 6:01 PM, Tom Lane wrote:

> David Jantzen <djantzen [at] ql2.com> writes:
>> Due to some historical idiosyncracies in our environment, we have a cust=
om 8.3.7 database installation built from source. We'd like to install dbl=
ink into this, however there are some problems with doing so:
>
>> 1) the 8.3.7 database was built on a CentOS 4 build box that has since g=
one away
>> 2) currently we have only 8.3.9 code built against CentOS 5
>> 3) the GCC compiler on CentOS 4 was quite old
>> 4) possible API changes in dblink between those versions
>
>> My question, how risky would it be to copy the dblink.so and .sql files =
from the CentOS 5 compilation of Postgres 8.3.9 over to the CentOS 4 compil=
ation of Postgres 8.3.7? If runtime errors result, how severe would they b=
e? I.e., would they take down a postgres backend or possibly the postmaste=
r daemon?
>
> I think you've got a fundamental problem that you'd better fix. If you
> are unable to rebuild the database from source then you are unable to
> update --- and you are already three minor versions behind and missing
> multiple security and crash-risk bug fixes. You're living on borrowed
> time, and NEED to reinstantiate your ability to build for that platform.
> Or move to a newer one.
>
> FWIW, you could probably get back a RHEL4 build environment pretty
> cheaply by setting up a suitable "mock" chroot on a recent Fedora
> version (installed on the same type of hardware).
>
> As for your direct question: no, I wouldn't count on that to work.
> RHEL4 and RHEL5 had different glibc versions didn't they?
>
> regards, tom lane


--
Sent via pgsql-admin mailing list (pgsql-admin [at] postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

--=-Neb/8aydcjxPXCaltbFW
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, 2010-03-15 at 19:25 -0700, David Jantzen wrote:
> We'll look at upgrading to 8.3.9 from an RPM.

FWIW 8.3.10 was released today.
--
Devrim G=C3=9CND=C3=9CZ
PostgreSQL Dan=C4=B1=C5=9Fman=C4=B1/Consultant, Red Hat Certified Engineer
PostgreSQL RPM Repository: http://yum.pgrpms.org
Community: devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org Twitter: http://twitter.com/devrimgunduz

--=-Neb/8aydcjxPXCaltbFW
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iEYEABECAAYFAkue7SQACgkQtl86P3SPfQ7aUACg05USL69hvgta4Ekx0NzY ri0e
L10AoN3S3IY/EzTJJrIjHM/h8eAkbJXI
=Fj2m
-----END PGP SIGNATURE-----

--=-Neb/8aydcjxPXCaltbFW--