Server rebooting 1 to 2 times weekly. Need help in checking cluster configuration.

Article:TECH143480  |  Created: 2010-11-04  |  Updated: 2011-09-13  |  Article URL http://www.symantec.com/docs/TECH143480
Article Type
Technical Solution


Environment

Issue



2 node CVM/CFS using 5.1RP1 on Linux RHEL5U3. Up and running. On 14th Oct, around 21:30 , 1 node got rebooted (no panic), customer wants to know the reason of reboot as this is happening 1-2 times in a week.


Error



Oct 14 21:24:39 newport2 kernel: nmupm[10095] trap divide error rip:804a35a rsp:ffaf4764 error:0 Oct 14 21:24:54 newport2 kernel: nmupm[10699] trap divide error rip:804a35a rsp:ff9435b4 error:0 Oct 14 21:25:08 newport2 kernel: program nmhs is using a deprecated SCSI ioctl, please convert it to SG_IO


Cause



Oracle patch issue


Solution



Check the following metalink - Subject: Error - program nmhs is using a deprecated SCSI ioctl, please convert it to SG_IO in /var/log/messages Doc ID: 566547.1 Type: PROBLEM Doc ID: 566547.1 Unpublished Bug 5897089 NMHS USES DEPRECATED SCSI IOCTL nmhs is used by the storage_report_metrics.pl which runs once every 24 hours by default. This script is querying the OS about various devices and mount points. This message is for information only and can be ignored. It is an indication that a future release of SCSI might not support this ioctl call. However, due to the fact that it floods the /var/log/messages file, a fix has been provided. Solution Patch 5897089 is available via My Oracle Support to address this issue.




Article URL http://www.symantec.com/docs/TECH143480


Terms of use for this information are found in Legal Notices