IBM Support

On-demand CPU frequency scaling hang during Red Hat Enterprise Linux 5 shutdown - Servers

Troubleshooting


Problem

394134, cogent:7251, cpu, d/t4190, d/t4192, d/t4367, hat, linux, processor, prog, program, red, software, system x, unclassified,

Resolving The Problem

Source

RETAIN tip: H192210

Symptom

In rare cases, the system hangs with the following CPU speed message during system shutdown:

Disabling ondemand CPU frequency scaling
Affected configurations

The system may be any of the following IBM servers:

  • System x3200 M2, type 4367, any model
  • System x3250 M2, type 4190, any model
  • System x3350, type 4192, any model

This tip is not option specific.

The system is configured with at least one of the following:

  • Red Hat Enterprise Linux 5, any update
  • RedHat Enterprise Linux 5.1, any update

Note: This does not imply that the network operating system will work under all combinations of hardware and software.

Please see the compatibility page for more information:

http://www.ibm.com/servers/eserver/serverproven/compat/us/

The CPU speed utility is affected.

Solution

This is a known issue in RHEL5 U1.

This behavior will be corrected in a future RHEL 5 release.

The Target date for this release is scheduled for Second Quarter 2008.

The file will be available from the IBM System x Support web site at the following URL:

http://www.ibm.com/systems/support/
Workaround

The user can select either workaround below to disable the CPU speed service in RHEL5 Update 1:

  1. Type the command 'chkconfig --level 12345 cpuspeed off' to disable the CPU speed service in all run levels.
  2. Type the command 'setup' and select System Services to unmark the CPU speed service. This affects only run level 5.
Additional information

The CPU speed service will enable / disable the cpufreq drivers, such as, cpufreq_ondemand and cpufreq_userspace during the system start up and shutdown.

The reason the issue occurs in RHEL5 U1 is that its synchronization mechanisms, such as, lock_cpu_hotplug and dbs_mutexare, are unstable in the cpufreq drivers, so that it may cause the system deadlock when the cpufreqs driver are under extreme loading conditions.

In our test, this issue occurred at random and was found only once during 1000 times of the system reboot test. Red Hat has planned to fix this issue in their future operation system update.

Document Location

Worldwide

Operating System

System x:Operating system independent / None

[{"Type":"HW","Business Unit":{"code":"BU016","label":"Multiple Vendor Support"},"Product":{"code":"HW334","label":"System x->System x3200 M2"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Line of Business":{"code":"","label":""}},{"Type":"HW","Business Unit":{"code":"BU016","label":"Multiple Vendor Support"},"Product":{"code":"HW335","label":"System x->System x3250 M2"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Line of Business":{"code":"","label":""}},{"Type":"HW","Business Unit":{"code":"BU016","label":"Multiple Vendor Support"},"Product":{"code":"HW336","label":"System x->System x3350"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
29 January 2019

UID

ibm1MIGR-5073625