intermitent looping redirect with blackboard

[Cantor, Scott]

On 10/18/12 11:38 AM, "caleb racey" <caleb.racey at newcastle.ac.uk> wrote:

>Out of 6000 logins to the idp coming from  blackboard today the logs show
>roughly 600 getting caught up in a looping redirects. Our blackboard
>landscape is 5 front end servers behind a load balancer (A10), sessions
>on the blackboard SPs  are being written  to and read from a shared
>database so they should all have access to the same shib session
>information, we have one live IDP.
>
>Has anyone experienced this kind of problem before and have tips of what
>to look for to avoid the issue?

I've not seen it with a shared cache, but there's not much you can do
without reproducing the problem and identifying the log information that
indicates why the session's not valid.

Without a shared cache, it's been seen here with NetScalers because the
cookie-based affinity option seems to not work for some clients.

I assume by shared database you mean the SP cache, not Blackboard. My
guess is the plugin's failing, but there should be something in the log
about it when the loops occur. You could turn up logging and get the
storage service category turned up to pick up the SQL and so forth.

-- Scott