Re: spontaneous node reboot

Metalink has a note that covers how to troubleshoot CRS reboots. The step by step troubleshooting process will help you track it down, or at least tell you what files you need to package up for support to track it down.

Note:265769.1

Subject: 	10g RAC: Troubleshooting CRS Reboots
Type: 	TROUBLESHOOTING
Status: 	PUBLISHED

Content Type: 	TEXT/X-HTML
Creation Date: 	16-MAR-2004
Last Revision Date: 	14-APR-2005

PURPOSE

To provide information on how to troubleshoot CRS reboots.

SCOPE & APPLICATION

This document is intended for DBA's and support analysts experiencing CRS reboots.

.....

Job

Billy wrote:
> johnle0701_at_yahoo.ca wrote:
> > We have Oracle RAC 10g Release 1 (10.1.0.2) installed
> > on two Solaris boxes (SPARC Solaris 9).
> >
> > Oracle RAC works well for about one working day.
> > Then for some reason, the nodes spontaneously reboot
> > overnight.
>
> There can be a number of reasons. E.g. when communication is lost on
> the interconnect by a node, it will feel left out of things, sulk, and
> reboot. ;-)
>
> Another reason is hardware. We had a case where cluster nodes just
> haphazzardly rebooted. Not a single char in any log file as to why. The
> cause was numerous screwups by "professional support & services" who
> installed the hardware without reading the friggen installation
> documentation.. like putting high speed PCI cards in slow speed shared
> PCI bus slots, mounted the units in the rack to close to one another
> causing them to run hot, and so on.
>
> So it can be literally anything. Bad/dirty power to the unit. Heat. H/w
> problems. Network problems. O/S problems. CRS issues. Etc. Etc.
>
> --
> Billy
Received on Tue Jun 28 2005 - 08:32:10 CDT