Am 06.06.2015 um 02:59 schrieb Justin Swanhart:
It was Linux's fault, not MySQL: https://bugzilla.redhat.com/show_bug.cgi?id=479765
this was *not* the 100% CPU bug i spoke about
This bug is about a specific leap-second deadlock that was fixed years ago. Please stop commenting in this bug about unrelated leap-second problems that occurred in 2012.
It also affected Java. http://serverfault.com/questions/403732/anyone-else-experiencing-high-rates-...
https://blog.mozilla.org/it/2012/06/30/mysql-and-the-leap-second-high-cpu-an...
So don't blame MySQL for being at the mercy of the OS.
when i have mail, web, file, dhcp, dns and what not servers all running fine at the mercy of the same OS i blame the one software running in circles
On Jun 5, 2015, at 3:47 PM, Reindl Harald <h.reindl@thelounge.net <mailto:h.reindl@thelounge.net>> wrote:
Am 05.06.2015 um 18:25 schrieb Justin Swanhart:
Wasn't that the result of a kernel bug?
well, when i have 30 servers running a recent kernel and every service except mysqld which starts to consume 100% CPU is running fine guess where the problem was - and frankly the only reason why the whole infrastructure did not went down was the availibility of 40 GHz so that public services still appeared to run fine, mysqld also responded "normal" at that time
On Jun 5, 2015, at 2:38 AM, Reindl Harald <h.reindl@thelounge.net <mailto:h.reindl@thelounge.net>> wrote:
can't speak for MariaDB, in case of MySQL the last leap second leaded in *each and every* mysqld running with 100% CPU on a dozens of machines
Am 05.06.2015 um 02:43 schrieb Marco Nicosia: I'm writing to inquire if I should prepare any of our MariaDB/Galera users for the upcoming leap second. According to http://www.timeanddate.com/time/leapseconds.html the next leap second will be added on June 30, 2015 at 23:59:60 UTC.
I searched a bit for "leap second mariadb" but didn't find anything recent.
Most of our users are on MariaDB 10.0.13 and Galera 25.2.9 or subsequent.
Have anyone tested both MariaDB and Galera to ensure that there are versions which are resilient to the introduction of a leap second? If so, is there documentation of versions which are known vulnerable to or proof against it?