Leap second problem causing high cpu after midnight 1st July


#1

I wonder if any non-managed BM servers (like me) have experienced this issue - high cpu on symbiosis-apache threads / java affecting mysql db backups, php/mysql setups (too many connections errors)?

The required fix is a snip - see

Apparantly BM are aware


#2

I didn’t notice any problem on Bytemark VMs, but I tend to run ntpd anyway. I have seen high load from Oracle and Java (Caucho Resin) processes, but that may be unrelated—for a start, it’s not unusual for Oracle and Java to consume lots of resources. In any case after a restart of these the systems had a more normal load.

The H has an article too:
http://www.h-online.com/open/news/item/Leap-second-bug-in-Linux-wastes-electricity-1631462.html


#3

my dedicated host also runs ntpd but still had the high cpu usage (on symbiosis-apache threads as well as java) following the leap second change. If anything it seems to me that ntpd update caused the problem since there’s no other way the leap second step could have been applied to the machine.