Unsuccessful read of Session

[Aaron Cargo]

Spot on, looks like shibd is recycling itself non-stop.

Before I go chasing that down, I wanted to make sure that this entry in the
shibd.log is unrelated:

INFO Shibboleth.Listener [240]: detected socket closure, shutting down
worker thread


Thanks!

Aaron Cargo
*Senior Web Developer*
Seton Hill University
acargo at setonhill.edu
(724) 552-4386



On Thu, May 12, 2016 at 5:03 PM, Cantor, Scott <cantor.2 at osu.edu> wrote:

> > We're using the same basic configuration for this SP as others in
> > our environment, but have logs full of the following:
> >
> > WARN Shibboleth.SessionCache [63361]: unsuccessful read of session (ID:
> > _42856448f2c3893bf74d8aba239859ed), caches out of sync?
>
> The bug Peter referenced is not referring to this getting logged when the
> sessions happen to have expired, otherwise it means what it says, they're
> not there anymore.
>
> > 1. Ensured time was correct and in sync on IDP and SP
>
> Won't matter here.
>
> > 2. Verified an appropriately long timeout setting in shibboleth2.xml
> (it's
> > consistent with other, working SPs in our environment).
>
> Rules out premature timeouts I guess.
>
> > 3. Verified that the default IDP NotOnOrAfter settings have not been
> > changed.
>
> It would be SessionNotOnOrAfter that's involved, not that setting.
>
> > I'm rather at a loss. I should note that this particular SP is behind a
> load
> > balancer, but when the issues were reported we removed all but a single
> > instance from the LB and the issue persists.
>
> Well, the sessions aren't there. If you're using a local shibd in the
> usual mode on the same box with the web server with no clustering, then I
> really have no explanation for you other than shibd restarting on you
> constantly, causing the data to disappear. Any sign of that?
>
> -- Scott
>
>
> --
> To unsubscribe from this list send an email to
> users-unsubscribe at shibboleth.net
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://shibboleth.net/pipermail/users/attachments/20160512/a8986d51/attachment.html>