Leap second tonight

Mailman List Mirror (Read Only)
1

bash-2.05b# date
Thu Jan 1 00:59:58 CET 2009
bash-2.05b# date
Thu Jan 1 00:59:59 CET 2009
bash-2.05b# date
Thu Jan 1 00:59:60 CET 2009
bash-2.05b# date
Thu Jan 1 01:00:00 CET 2009
bash-2.05b# date
Thu Jan 1 01:00:01 CET 2009
bash-2.05b#

-P

3

At which point my Solaris 10 v490's reboot in unison, lovely.

Anyone else see anything interesting?

-wil

4

My Solaris 10 boxes are all happy (and did not reboot). I monitor NTP
on a number
of devices, including one router. The router was off by one second for
a while, but
is OK after an hour. Everything else was fine immediately.

In 2005, our CDMA clock got the leap second between 15:08 and 15:38
EST creating
some issues due to disagreement with the (too few) GPS clocks.

Jon

5

Jon Meek wrote:

My Solaris 10 boxes are all happy (and did not reboot). I monitor NTP
on a number
of devices, including one router. The router was off by one second for
a while, but
is OK after an hour. Everything else was fine immediately.

In 2005, our CDMA clock got the leap second between 15:08 and 15:38
EST creating
some issues due to disagreement with the (too few) GPS clocks.

Jon

At which point my Solaris 10 v490's reboot in unison, lovely.

Anyone else see anything interesting?

-wil

I run a bunch of Slackware Linux boxes of varying versions. As best as I can tell, at or around 00:00 UTC all of my Slackware 12.0 boxes crashed with a kernel panic. I don't think it is ntpd because it is the same version as on 12.1 boxes (4.2.4p0) that did not crash. It may be the kernel: 2.6.21.5

Anyone else experience similar or was this coincidental and I have other issues...

Steve

6

Solaris? Or ZuneOS? (See
http://www.nytimes.com/2009/01/01/technology/personaltech/01zune.html)

    --Steve Bellovin, http://www.cs.columbia.edu/~smb

7

I had a couple of Oracle servers (Solaris 10) reboot a couple of minutes
just before the leap second. All my other Solaris 10 boxes appear to have
stayed up fine.

Simon

8

Have either of you determined if this was a OS reboot and not a bios reset?

-Jim P.

9

I've been trawling through all the logfiles I can find on the box, and I see
normal entries up until 23:59:xx, and then the next entry is stuff restarting.
Could well be a BIOS/LOM reset, but it's odd that the only two boxes affected
were my Oracle servers.

Simon

10

All of my Solaris 10 boxes stayed up with the exception of the Oracle 10g RAC boxes.

db1:~ wschultz$ uname -a
SunOS db1 5.10 Generic_137111-01 sun4u sparc SUNW,Sun-Fire-V490

A friend of mine had his RAC boxes reboot as well, similar configuration. I've poured through the logs and see normal activity until the reboot, nothing suspicious and no reason for the reboot. Seems to be specific to Solaris 10 running RAC on this end.

-wil

11

My Oracle boxes that rebooted were running RAC (version 10G R2), too. Another
Solaris 10 box running the same version of Oracle, but not RAC, did not reboot.

Looks rather like an Oracle 10 RAC bug.

Simon

12

I've got a 10g RAC cluster with T2000s and 2900s that didn't reboot. I
can dig up specific release levels, but I think we're at 10.2.0.4

-j

13

Once upon a time, Steven Saner <ssaner@hubris.net> said:

I run a bunch of Slackware Linux boxes of varying versions. As best as I
can tell, at or around 00:00 UTC all of my Slackware 12.0 boxes crashed
with a kernel panic. I don't think it is ntpd because it is the same
version as on 12.1 boxes (4.2.4p0) that did not crash. It may be the
kernel: 2.6.21.5

Anyone else experience similar or was this coincidental and I have other
issues...

I had one (out of many, including about a half dozen identically
configured) Red Hat Enterprise Linux 4 systems hang at the leap second.

There have been some messages on the NTP list referencing posts on a
Debian list about leap second crashes, and there's a post on /. about a
similar problem with Fedora 8.

14

Steven Saner wrote:

Jon Meek wrote:

My Solaris 10 boxes are all happy (and did not reboot). I monitor NTP
on a number
of devices, including one router. The router was off by one second for
a while, but
is OK after an hour. Everything else was fine immediately.

In 2005, our CDMA clock got the leap second between 15:08 and 15:38
EST creating
some issues due to disagreement with the (too few) GPS clocks.

Jon

At which point my Solaris 10 v490's reboot in unison, lovely.

Anyone else see anything interesting?

-wil

I run a bunch of Slackware Linux boxes of varying versions. As best as I
can tell, at or around 00:00 UTC all of my Slackware 12.0 boxes crashed
with a kernel panic. I don't think it is ntpd because it is the same
version as on 12.1 boxes (4.2.4p0) that did not crash. It may be the
kernel: 2.6.21.5

Anyone else experience similar or was this coincidental and I have other
issues...

Steve

Yep. I have a few Slack 12 boxes lockup. Digging around, it looks to be
a issue with pre 2.6.21.5 kernels.

-Eddie

15

It's a known bug in Oracle 10. When the time is set backwards the system reboots. I don't have the bug id at hand but there is one, and a patch.

Either patch or don't run NTP on Oracle servers.