Hi,
We use RADIUS authentication at $WORK, and it has one major flaw (well,
two, but I already fixed the other one this week): it only supports
specifying a single server, which as we might know, is bad for high
availability. So I'm developing a patch to fix this issue, but I'm not
exactly sure what the configuration should look like. I see multiple
options, but the one I like the best is the following:
Add two new HBA configuration options: radiusfallbackservers and
radiusfallbackports; both lists parsed with SplitIdentifierString (à la
listen_addresses). If radiusfallbackservers is set, it's the list of
servers to try after "radiusserver" has timed out (but ONLY if it timed
out, obviously). If radiusfallbackports is set, it must be of equal
length with radiusfallbackservers, and the corresponding port entry is
used for each fallback server. If it's not set, radiusport is used as
the port for all fallback servers, or finally the default port (1812) if
no port options have been specified in pg_hba.conf.
This interface is both flexible, simple to use in the common case (YMMV)
and fully backwards compatible.
I'm planning two submit two patches: one to refactor CheckRADIUSAuth() a
bit to make it easier to try the authentication against more than one
server, and then another one to add the two new configuration options.
Does anyone have a problem with this plan? Want to comment on the
interface or get anything else off their chest?
.m