upgrade 5.0.5 to 5.0.6 pre auth failures

[phingers]

Upgraded to 5.0.6 last night, and now all my code that uses pre-auths dont work.

gdpak domain.com says there already is a key for the domain.

Do I need to wipe them out and re-generate?

They should have worked fine after the upgrade right?

Thoughts?

[webman]

Hi

I did an upgrade to 5.0.6 too and the preauth stuff still works; so I don't think it's by design that you have to re-generate yor preauth stuff.

You can get the current preauth key by running this:

zmprov gd example.com|grep zimbraPreAuthKey

This should match what you have in your applications. If not, then you can set it again (to what you already have) using something like:

zmprov md example.com zimbraPreAuthKey {my-existing-key}

[phingers]

So I have the wrong error I guess...

What would cause this after the upgrade?

=============================

HTTP ERROR: 400

authentication failed for user@domain.com

RequestURI=/service/preauth

Powered by Jetty://

[phingers]

2008-05-27 11:44:49,389 WARN [btpool0-22] [] security - cmd=PreAuth; account=user@domain.com; admin=false; error=authentication failed for user@domain.com, preauth timestamp is too old, server time: Tue May 27 11:44:49 CDT 2008, preauth timestamp: Tue May 27 11:49:55 CDT 2008;

[webman]

Looks like there's a huge time difference (more than five minutes) between Zimbra and the computer calculating the preauth. Try adjusting the time so they're both the same (use NTP?) and try again.

[phingers]

The two server date/times were off by more than 5 mins. I corrected by running:

/usr/sbin/ntpdate -b -s us.pool.ntp.org

Manually on the zimbra server, since that one seemed to be slower. Weird how that happened though.

[webman]

Cool

Sometimes this happens after a reboot. Stick that ntp command in root's daily cron and you should be OK though