My impression was that update manager should make a new BE, then run update on the new BE and set the new BE to be used on the next reboot. This doesn't seem to happen.
I recently tried updating to the latest development build for 2010.02, feeling safe that if there were problems I could revert to my current boot environment. The update failed - my system wouldn't start (more on that elsewhere) so I reverted to my previous BE. On doing so I found that my system was trying to use a "wrong" version of MySQL (I suffered from the bug in 5.1.30 (http://bugs.mysql.com/bug.php?id=41710) so had rolled my own. I needed mysql running so could spend a lot of time trying to work out what had happened. This is what I could see. - the mysql version running was 5.1.30 (I believed 5.1.37 should be in SVN 127 based on http://www.opensolaris.org/jive/click.jspa?searchID=2585408&messageID=418214 but was mistaken) - files dated after the update were in mysql bin directories. I solved the problem by creating a new BE using one of timeslider's automated snapshots from prior to my update attempt. But this seems to defeat the purpose of the Update Manager. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
