On Sunday, 1. July 2012. 15.06.34 Reindl Harald wrote: > Am 01.07.2012 14:47, schrieb Tom Horsley: > > They really ought to switch to a new version of NTP > > protocol that is just like the old one, but works from > > TAI and sends a database of leap second info around as > > well so computers can translate TAI into UTC. > > in case of a bug whatever new protocol will not help > this seems more likely to be a kernel bug because > there were many apps on different machines affected > > * BIND > * MySQL (hardly!) > * Apache Traffic Server > * Firefox > * Thunderbird > * who knows what else > > see also: > http://bugs.mysql.com/bug.php?id=65778 > > i recently restarted ALL computers in our production envirnonment > up to any VMware ESXi Host, SAN-Storage-Controllers (Managment/Storage) > even up to my Android-Phone for security after woke up and saw tons of > alarms about high CPU usage on the whole infrastructure caused > mostly by 15 mysqld instances (and saw the same at home in VMware-Guests > and host) > > yes, my first guess to the mysql-list was that i think it > has something to do with teh leap-second last night
I've got bitten by this as well. It was a kernel bug, mishandling the leap second (AFAIU, it left the door open for some race condition to happen or not happen, and if it happens...). See http://lkml.indiana.edu/hypermail/linux/kernel/1203.1/04598.html Of course, the workaround is to reset the date or reboot the machine, whichever is easier. ;-) Best, :-) Marko -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
