PostgreSQL 7.3.7 has a risk of data loss explained here: http://archives.postgresql.org/pgsql-hackers/2004-10/msg00464.php In addition there are two minor security issues reported to us by iDEFENSE; these were judged too minor to warrant assigning CVE names, but they should be fixed anyway. Both are escalation-of-privilege issues that might allow the postgres user to obtain root. The first is that "pg_ctl start" executes a shell command using text obtained from a postgres- owned configuration file. This is only a minor risk because "pg_ctl start" should never be run as root anyway; but possibly a sysadmin unfamiliar with Postgres could be tricked into doing it. The appropriate fix is to disallow pg_ctl from running as root. The second issue is that the postgresql init script executes a "chown" as root without sufficient care about what is being chown'ed; a nefarious postgres user could install a symlink to cause any desired file to be chown'd to postgres. (This was judged a minor issue only because installation of the symlink would break the Postgres installation, and thus would be rather unlikely to go undetected until next system reboot.) The first two issues are solved in the upstream Postgres 7.3.8 release, which we should upgrade to. The init-script issue we need to fix locally. Version-Release number of selected component (if applicable): postgresql-7.3.7-3 How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
This is repaired in rh-postgresql-7.3.8.1.
An errata has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2004-489.html