Two of my systems injected leap seconds at the end of Aug and Sep.
A couple of others didn't.

I thought I saw some discussion of this, but I can't find it. Has
anybody else seen this? Was I dreaming about seeing something here?

Is the bug in the Linux kernel for not waiting until December, or
should ntpd wait until December before telling the kernel?

I think both systems are/were running a recent ntp-dev on a reasonably recent
Linux (2.6) kernel.

I'm watching (noselect) both systems from a third box that didn't have
this glitch.

One did what I expect. It's clock was off eby a second. After about 20
minutes it stepped back.

I haven't sorted out what the other one did yet.

--
These are my opinions, not necessarily my employer's. I hate spam.