Shibboleth SP getting progressively slower

[Cantor, Scott]

On 1/20/12 1:10 PM, "Steve Thorpe" <thorpe at mailbox.mcnc.org> wrote:
>
>Out of curiosity, what is the default duration after which the reverse
>index would expire?

It expires when the cache entry for the session does. Or in this case when
the last entry containing the affected NameID expires. The specific rules
on that are very complicated because of the mix of deprecated and current
options controlling cache timeouts. With newer defaults it's designed to
be essentially at the same time as the session would expire.

>If a monitoring frequency could be adjusted to be a
>bit longer, then the monitoring case presumably wouldn't encounter this
>issue even when using a fixed NameID.

No, but you probably wouldn't want to monitor as infrequently as your
session lifetime.

-- Scott