Compute Node Reboot with ZERO DIVIDE AT FIND_BUSIEST

Asset ID:	1-72-1615373.1
Update Date:	2017-01-30
Keywords:

Solution Type Problem Resolution Sure

Solution 1615373.1 : Compute Node Reboot with ZERO DIVIDE AT FIND_BUSIEST_GROUP

Applies to:

Exadata X3-2 Hardware - Version All Versions to All Versions [Release All Releases]
Information in this document applies to any platform.

Symptoms

Compute node was restarted without any errors in message, CRS, CSSD. No load issues are seen in oswatcher logs.

Message file shows:

Nov 13 20:28:34 exakdbadm02 auditd[5220]: Audit daemon rotating log files
Nov 13 20:28:34 exakdbadm02 auditd[5220]: Audit daemon rotating log files
Nov 13 23:10:47 exakdbadm02 syslogd 1.4.1: restart. <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
Nov 13 23:10:47 exakdbadm02 kernel: klogd 1.4.1, log source = /proc/kmsg started.
Nov 13 23:10:47 exakdbadm02 kernel: Initializing cgroup subsys cpuset

Changes

ILOM snapshot shows error:

RIP [<ffffffff81053c35>] find_busiest_group+0x7ba/0x88e
RSP <ffff882ed9645948>
---[ end trace 44010a4c56e3c3b8 ]---
divide error: 0000 [#2]
Kernel panic - not syncing: Fatal exception <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

Cause

This is exact issue due to Bug 16483310 - ZERO DIVIDE AT FIND_BUSIEST_GROUP+0X447/0X90E

Solution

Action Plan:
=========

The bug is fixed in 2.6.32-400.24.1.el5uek.

Upgrade to at least above kernel version and it is advisable to upgrade to latest version 2.6.32-400.34.1.el5uek as described in Doc ID 1613511.1

For exadata customers it is advisable to Upgrade to Exadata image version 11.2.3.2.2 which contains the Bug fix and it updates the USB as well

Attachments

This solution has no attachment