System freezes after more than 417 days uptime

This document (7014192) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 10 Service Pack 4 LTSS for IBM System z
SUSE Linux Enterprise Server 11 Service Pack 1 LTSS for IBM System z
SUSE Linux Enterprise Server 11 Service Pack 2 for IBM System z

Situation

After more than 417 days of uptime the kernel floods the syslog with "update_cpu_power: cpu_power = ..." messages and the machine becomes unresponsive.

Resolution

For SUSE Linux Enterprise Server 10 SP4 the issue has been fixed with a maintenance update of the Linux kernel to version 2.6.16.60-0.101.1.

For SUSE Linux Enterprise Server 11 SP2 the issue has been fixed with a maintenance update of the Linux kernel to version 3.0.74-0.6.6.2.

For Long Term Service Pack Support customers of SUSE Linux Enterprise Server 11 SP1, please contact SUSE Linux Support for a PTF kernel until a kernel update has been published.

As a workaround, schedule a reboot of the Linux system before reaching an uptime of 417 days.

Cause

To convert the TOD clock value to nano seconds, the Linux kernel multiplies the value by 125 and then divides it by 512 within sched_clock(). This triggers an overflow after approximately 417 days that leads to erratic behavior of kernel subsystems that rely on monotonic sched_clock() results.

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.