Sun Fire[TM] 12K/15K/E20K/E25K: Domain Reboots if the Open Boot PROM Boot Device Is Incorrect

Asset ID:	1-72-1005038.1
Update Date:	2016-08-17
Keywords:

Solution Type Problem Resolution Sure

Solution 1005038.1 : Sun Fire[TM] 12K/15K/E20K/E25K: Domain Reboots if the Open Boot PROM Boot Device Is Incorrect - dxs disconnected message

Applies to:

Sun Fire 15K Server - Version All Versions and later
Sun Fire E20K Server - Version All Versions and later
Sun Fire 12K Server - Version All Versions and later
Sun Fire E25K Server - Version All Versions and later
All Platforms
***Checked for relevance on 23-Apr-2011***

Symptoms

A domain on a Sun Fire[TM] 12K/15K/E25K platform reboots if the domain cannot boot off the Open Boot PROM boot device, either because it is set incorrectly or because of a hardware failure.

Cause

For example, if a user modifies the boot device of a domain incorrectly and then tries to boot the domain, the domain will be rebooted automatically. The user is then disconnected from the domain console. Messages similar to the following will appear on the console:

{160} ok boot /pci@17c,700000/pci@1/scsi@2/sd@0,0:a
Boot device: /pci@17c,700000/pci@1/scsi@2/sd@0,0:a  File and args:
Can't locate boot device
{160} ok The dxs disconnected.
Trying to connect...
( The system is running POST )

Messages similar to the following should appear in the domain messages file on the System Controller at:

$SMSVAR/SMS/adm/--->domain_id<---/messages:
Feb 28 10:32:40 2003 xc9p13-sc1 dsmd[7207]-M(): [2536 1359068189896205 NOTICE DomainsPatrol.cc 664] Reset domain M request received, restarting domain.
Feb 28 10:34:35 2003 xc9p13-sc1 dsmd[7207]-M(): [2537 1359183873470286 NOTICE DomainsPatrol.cc 747] Domain M OBP failed, rebooting domain.
Feb 28 10:34:35 2003 xc9p13-sc1 dsmd[7207]-M(): [0 1359183874528188 NOTICE SysControl.cc 485] Since domain failure occurred at OBP, set boot command to 
forced-no-boot and recover domain to OBP only
Feb 28 10:34:55 2003 xc9p13-sc1 dxs[22543]-M(): [4916 1359204021832799 ERR ConsoleService.cc 845] DXS - Iosram getFlag failed or busy in CONI. busy flag=1,rc=0

Solution

This is a feature of the Sun System Management Software (SMS) and the system is performing as designed. It is actually the domain status monitoring daemon (DSMD) controlled by SMS that initiates the reboot
action. DSMD has no way to determine whether the cause is due to a user's typographical error or due to a true error. For example, the error reported by Open Boot PROM could be a true error for those occasions when the boot disk is powered down.

To discuss this information further with Oracle experts and industry peers, we encourage you to review, join or start a discussion in an appropriate
My Oracle Support Community - Oracle Sun Technologies Community.

Internal Comments
This is behavior will not be changed. See internal Bug ID 4825660
See also <Document 1004721.1>

Can't locate boot device, OBP failed, rebooting domain, 12K, 15K, E20K, E25K, boot-device, incorrect, The dxs disconnected, starcat
Previously Published As 76798

Attachments

This solution has no attachment