Skip to Content.
Sympa Menu

shibboleth-dev - RE: Clustering an SP?

Subject: Shibboleth Developers

List archive

RE: Clustering an SP?


Chronological Thread 
  • From: "Scott Cantor" <>
  • To: <>
  • Subject: RE: Clustering an SP?
  • Date: Tue, 29 Nov 2005 00:20:24 -0500
  • Organization: The Ohio State University

> Is there any support in the works, or otherwise planned, to
> support SP session cache over a cluster of SP systems?
> This is for a unix, apache sp.

I think the only thing planned is what's there now, an API to do it. A
non-embedded MySQL plugin is probably the LCD, and it's a joke from
everything I've seen. Googling for MySQL + session cache you find non-stop
mentions of table corruption due to "lots of writes". Umm, ok. Postgres is
free, but much less widely used (and I know nothing about it).

Lexis Nexis offered an Oracle plugin for 1.2, but it hasn't been packaged
for contribution, and it would need to be updated for 1.3 and beyond.
Buffalo also has an Oracle plugin, also for 1.2 I think.

I don't see any way I could produce a portable plugin outside of Java. I
suppose a plugin could be written to some proprietary API, similar to ODBC,
but ODBC isn't all that prevalent on Unix. Plus there's that problem of just
adding a single point of failure back into the cluster.

The JBoss stuff Chad has been working on seems like a cool possibility for
that SP, eventually. The 2.0 design might lend itself to optionally using
Java for significant portions of the Apache SP as well, so there might be
ways to leverage that in the future.

-- Scott



  • Clustering an SP?, Jim Fox, 11/28/2005
    • RE: Clustering an SP?, Scott Cantor, 11/29/2005

Archive powered by MHonArc 2.6.16.

Top of Page