Metadata validity interval and fun with timezones

[Paul Hethmon]

Doesn't Daylight savings time end next weekend? Meaning at 2AM on the 3rd?

On 10/31/13 11:28 AM, "Wessel, Keith" <kwessel at illinois.edu> wrote:

>Hi, all,
>
>Our IDPs all started complaining about the metadata validUntil time being
>too far in the future this week. Our local metadata has a 7-day validUnti
>time, and the validityInterval in relying-party.xml is set to 7 days.
>Starting at exactly 2 AM on Sunday morning, it started complaining about
>the validity interval being between 45 and 60 minutes too long; we update
>local metadata every 15 mins.
>
>The time interval and the 2 AM Sunday thing seem to scream timezone data
>problem. Problem is we're on a recent Java release - 6.0U45. And it has
>the 2012i tzdata in it. I even updated our tzdata to the latest available
>from Oracle 2013g. 2013h came out last week. Oracle doesn't have an
>updater for it yet, but I'm doubtful that'll help.
>
>My work-around for now is to make my validity interval a smidge longer
>than the 7 days, and I'm fine with this. But I'm wondering. Can anyone
>think of something other than the timezone data itself that might be
>behind this? Everything else on the system is happy.
>
>Thanks,
>Keith
>
>--
>To unsubscribe from this list send an email to
>users-unsubscribe at shibboleth.net