'Leap second' snafu affects Oracle clustering tool

Cluster Ready Services software is rebooting in some cases because of second added to 2008. The second that time-tracking scientists added to 2008 to adjust for the slowing of the earth's rotation is causing problems with Oracle Corp.'s Cluster Ready Services (CRS) software, the vendor said in a support document issued Monday. But it added that a pair of fixes are available.

CRS is used in conjunction with Oracle's Real Application Clusters software, which enables a single Oracle database to be deployed on a group of servers, or "nodes," in order to provide fault tolerance and increased scalability.

The "leap second event" is causing CRS nodes to reboot, according to the Oracle document that details the problem. The affected database platforms are Versions 10.1.0.2 to 11.1.0.7 of the Oracle Server Enterprise Edition, running on 64-bit Sun Solaris servers with CRS and Oracle patch sets 10.2.0.1 to 11.1.0.7.

Coordinated Universal Time, the world's time standard, needs to be adjusted in order to account for the differences that develop between UTC and the international atomic clock because of Earth's rotational speed. The addition of leap seconds are handled by the International Earth Rotation and Reference Systems Service, either at the end of June or December. The IERS, as the organization is known, added a second to Dec. 31.

As a result, Network Time Protocol daemons "had to adjust time accordingly, and the CRS product stack has encountered problems resulting in node reboots," Oracle said in its support document. NTP is used to synchronize the clocks of computers and relies on UTC to provide reference times.

The reboots will occur on affected nodes only under two specific conditions, which were detailed by Oracle. The document also spells out two methods for fixing the issue, including the installation of available patches.

A spokesman for Oracle didn't immediately respond to a request for further comment.

The rebooting issue has prompted some discussions on multiple user forums and mailing lists in recent days.

"This begs the question -- how the heck do timekeepers and politicians get away with last-minute time changes?" one user wrote. "Surely there's some pushback from technology-related interest groups to try and get more than four weeks' warning?"

Other posters, however, pointed out that the IERS made its announcement regarding the plan to add the most recent leap second last July.

Oracle's disclosure follows the problem that cropped up last week with some of Microsoft Corp.'s Zune media players. The affected devices froze up and wouldn't work on Wednesday, a snafu that Microsoft attributed to a bug in their internal clock drivers. The bug became an issue because 2008 was a leap year, Microsoft said.

Author: Chris Kanaracus @ www.computerworld.com

Read more ...