Thread: RPMS on PostgreSQL.org updated.
I have updated the RPM distribution on postgresql.org today. The last two releases, 6.5.3-2 and -3, in both locale and non-locale versions, are available. I'm keeping two releases rotating online in case serious problems are found in the latest release that were not in the next-to-latest release. The directory structure for the RPMS has been changed somewhat, with the RPMs now being under /pub/bindist/RPM, rather than being in /pub/RPM and /pub/SRPMS, although those two locations have been symlinked into the appropriate directories under /pub/bindist/RPM so that preexisting links wouldn't be broken. Many thanks to Marc and Jeff for making the update possible. Latest release: postgresql-6.5.3-3 and 6.5.3-3nl. Please read /pub/bindist/RPM/README and README.rpm for more information. -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
> Latest release: postgresql-6.5.3-3 and 6.5.3-3nl. > Please read /pub/bindist/RPM/README and README.rpm for more > information. So should I blow away all of the other RPM directories on the FTP site?? - Thomas -- Thomas Lockhart lockhart@alumni.caltech.edu South Pasadena, California
Thomas Lockhart wrote: > > > Latest release: postgresql-6.5.3-3 and 6.5.3-3nl. > > Please read /pub/bindist/RPM/README and README.rpm for more > > information. > > So should I blow away all of the other RPM directories on the FTP > site?? You can, if you wish. Jeff gave me the bindist/RPM directory, and symlinked the RPMS and SRPMS tree of yours into the bindist/RPM/RPMS and bindist/RPM/SRPMS directories, respectively. The links on the download and news pages will have to be changed when you do so -- which is why I suggested to Jeff that he symlink, so that links wouldn't be broken. Future download links for the RPMs should point to bindist/RPM, IMO. I have already found some other annoyances with the RPMs -- on upgrading the server subpackage, a chkconfig --del is performed -- but a corresponding chkconfig --add is not performed, thus removing postgresql from being autostarted. I think a change in the behavior to leave well enough alone if upgrading, and do the chkconfig --del only if this is an uninstall would be appropriate. I got bit with this on my production server -- not fun. -- Lamar Owen WGCR Internet Radio 1 Peter 4:11