Oddness with AWS hosted IdP(s)
Norman Bodnar
bodnarn at gmail.com
Thu Jul 19 14:45:05 EDT 2018
Hi Scott, I agree, it's out of date... but for a lot of us balancing out a
couple dozen other systems and OSs, unless there is a security issue,
3/17/2017 version of Jetty doesn't sound "far out of date" to many of us --
but that's on me, I paid that price.
To wrap up, I found this on the jetty bug list...
The problem looks to be that that RollOverFileOutputStream.toMidnight
function figures out what day it is in the local time zone, but then
computes the delay based on the timezone passed in. In this case, any time
between 17:00 and 24:00 on day x in PDT computes midnight as midnight at
the start of day x in GMT, and nextMidnight at midnight on day x+1 in GMT.
However, the current System.currentTimeMillis() time is already past
midnight on day x+1 in GMT as soon as the time is past 17:00 on day x.
Hence, the negative delay in midnight.toInstant().toEpochMilli() -
System.currentTimeMillis().
>From <https://github.com/eclipse/jetty.project/issues/1507>
This may be the actual cause of our issues because we rebooted later in the
evening, something we've only done earlier in the day previously - and it
probably explains why it all miraculously worked after 1am.
On Thu, Jul 19, 2018 at 10:50 AM, Cantor, Scott <cantor.2 at osu.edu> wrote:
> > Thanks Scott. At present, are on Jetty 9.3.17.v20170317.
>
> Which basically answers your question, that's far out of date.
>
> -- Scott
>
> --
> For Consortium Member technical support, see https://wiki.shibboleth.net/
> confluence/x/coFAAg
> 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/20180719/f60cb57e/attachment.html>
More information about the users
mailing list