Shutting Down and Booting a SPARC Based System for Recovery Purposes

In the following instances, you must first shut down a system to analyze or troubleshoot booting and other system problems.

Troubleshoot error messages when the system boots.
Stop the system to attempt recovery.
Boot a system for recovery purposes.
Force a crash dump and reboot of the system.
Boot the system with the kernel debugger by using the kmdb command.

The procedures that follow describe how to safely shut down and then boot a SPARC based system for recovery purposes.

Stopping and Booting for System Recovery Purposes

You might need to boot the system for recovery purposes. Some of the more common error and recovery scenarios include the following:

Boot a system in single-user mode to resolve a minor problem, such as correcting the root shell entry in the /etc/passwd file or changing a NIS server.
Boot from the installation media or from an install server on the network to recover from a problem that is preventing the system from booting or to recover from a lost root password. Resolving a boot configuration problem by importing the root pool, mounting the BE, and fixing the problem.

On SPARC systems, the boot net:dhcp command replaces the boot net command that is used in Oracle Solaris 10 releases.

How to Stop a System for Recovery Purposes

Bring the system to ok PROM prompt by using the shutdown or init 0 command.
Synchronize the file systems.
```
ok sync
```
Type the appropriate boot command to start the boot process.
For more information, see the boot(1M) man page.

Verify that the system was booted to the specified run level.

# who -r
 .       run-level s  May  2 07:39     3      0  S

Example 9-1 Powering Off a Service Processor

If you are running Oracle Solaris 11 on an Oracle ILOM service processor, after shutting down the operating system, you must switch from the system console prompt to the service processor prompt. From there, you can stop the service processor, as shown in this example:

# shutdown -g0 -i0 -y
# svc.startd: The system is coming down. Please wait.
svc.startd: 91 system services are now being stopped.
Jun 12 19:46:57 wgs41-58 syslogd: going down on signal 15
svc.stard: The system is down.
syncing file systems...done
Program terminated
r)eboot o)k prompt, h)alt?
# o

ok #.
->

-> stop /SYS
Are you sure you want to stop /SYS (y/n)? y
Stopping /SYS
 
->

If you need to perform an immediate shutdown, use the stop -force -script /SYS command. Before you type this command, ensure that all data is saved.

Example 9-2 Powering On a Service Processor

The following example shows how to power on the server. You must first be logged in to Oracle ILOM. See http://download.oracle.com/docs/cd/E19166-01/E20792/z40002fe1296006.html#scrolltoc.

If you have a modular system, make sure you are logged into the desired server module.

-> start /SYS
Are you sure you want to start /SYS (y/n) ? y
Starting /SYS
 
->

If you do not want to be prompted for a confirmation, use the start -script /SYS command.

How to Boot in Single-User Mode to Resolve a Bad `root` Shell or Password Problem

Bring the system to the ok PROM prompt. See How to Stop a System for Recovery Purposes.
Boot the system in single-user mode.
```
ok boot -s
```
Correct the shell entry in the /etc/passwd file.
```
# vi /etc/password
```
Press control-d to reboot the system.

How to Boot From Media to Resolve an Unknown `root` Password

Boot the system from the Oracle Solaris media.
- Text installation – Boot from the install media or from the network, then select Option 3 Shell from the text installation screen.
- Automated installation – Use the following command to boot directly from an installation menu that allows you to exit to a shell.
```
ok boot net:dhcp
```
At the shell prompt, import the root pool.
```
# zpool import -f rpool
```
Create a mount point for the boot environment.
```
# mkdir /a
```

Mount the boot environment.

# beadm mount solaris-instance|bename /a

Set the TERM type.
```
# TERM=vt100
# export TERM
```
Carefully remove the unknown password entry.
```
# cd /a/etc
# vi shadow
# cd /
```
Note - You must change directories after this step.
Update the boot archive.
```
# bootadm update-archive -R /a
```
Unmount the boot environment.
```
# beadm umount be-name
```
Halt the system.
```
# halt
```

Reboot the system in single-user mode, and when prompted for the root password, press Return. For example:

ok boot -s
Boot device: /pci@780/pci@0/pci@9/scsi@0/disk@0,0:a File and args: -s
SunOS Release 5.11 Version 11.0 64-bit
Copyright (c) 1983, 2011, Oracle and/or its affiliates. All rights
reserved.
Booting to milestone "milestone/single-user:default".
Hostname: tardis.central
Requesting System Maintenance Mode
SINGLE USER MODE
Enter user name for system maintenance (control-d to bypass): root
Enter root password (control-d to bypass): <Press return>
single-user privilege assigned to root on /dev/console.
Entering System Maintenance Mode

Reset the root password.

root@system:~# passwd -r files root
New Password: xxxxxx
Re-enter new Password: xxxxxx
passwd: password successfully changed for root

Press control-d to reboot the system.

How to Boot a System Without Starting Any Services

If problems with starting services occur, sometimes a system hangs during the boot process. This procedure shows how to troubleshoot this problem.

Boot without starting any services.
This command instructs the svc.startd daemon to temporarily disable all services and start sulogin on the console.
```
ok boot -m milestone=none
```
Log in to the system and become the root role.
Enable all services.
```
# svcadm milestone all
```
When the boot process hangs, determine which services are not running and where the boot process is hanging.
```
# svcs -a
```
Check for error messages in the log files in /var/svc/log.
After fixing the problems, verify that all services have started.
1. Verify that all needed services are online.
```
# svcs -x
```
2. Verify that the console-login service dependencies are satisfied.
  This command verifies that the login process on the console will run.
```
# svcs -l system/console-login:default
```
Continue the normal boot process.

Forcing a Crash Dump and Reboot of a SPARC Based System

Forcing a crash dump and reboot of the system are sometimes necessary for troubleshooting purposes. The savecore feature is enabled by default.

For more information about system crash dumps, see Chapter 17, Managing System Crash Information (Tasks), in Oracle Solaris Administration: Common Tasks.

How to Force a Crash Dump and Reboot of the System

Use this procedure to force a crash dump of the system. The example that follows this procedure shows how to use the halt -d command to force a crash dump of the system. You will need to manually reboot the system after running this command.

Bring the system to the ok PROM prompt.
Synchronize the file systems and write the crash dump.
```
> n
ok sync
```
After the crash dump is written to disk, the system will continue to reboot.
Verify that the system boots to run level 3.
The login prompt is displayed when the boot process has finished successfully.
```
hostname console login:
```

Example 9-3 SPARC: Forcing a Crash Dump and Reboot of a System by Using the halt -d Command

This example shows how to force a crash dump and reboot of the system by using the halt -d and boot commands.

# halt -d
Jul 21 14:13:37 jupiter halt: halted by root

panic[cpu0]/thread=30001193b20: forced crash dump initiated at user request

000002a1008f7860 genunix:kadmin+438 (b4, 0, 0, 0, 5, 0)
  %l0-3: 0000000000000000 0000000000000000 0000000000000004 0000000000000004
  %l4-7: 00000000000003cc 0000000000000010 0000000000000004 0000000000000004
000002a1008f7920 genunix:uadmin+110 (5, 0, 0, 6d7000, ff00, 4)
  %l0-3: 0000030002216938 0000000000000000 0000000000000001 0000004237922872
  %l4-7: 000000423791e770 0000000000004102 0000030000449308 0000000000000005

syncing file systems... 1 1 done
dumping to /dev/dsk/c0t0d0s1, offset 107413504, content: kernel
100% done: 5339 pages dumped, compression ratio 2.68, dump succeeded
Program terminated
ok boot
Resetting ... 

.
.
Rebooting with command: boot
Boot device: /pci@1f,0/pci@1,1/ide@3/disk@0,0:a
File and args: kernel/sparcv9/unix
configuring IPv4 interfaces: hme0.
add net default: gateway 172.20.27.248
Hostname: jupiter
The system is coming up.  Please wait.
NIS domain name is example.com
.
.
.
System dump time: Wed Jul 21 14:13:41 2010
Jul 21 14:15:23 jupiter savecore: saving system crash dump
in /var/crash/jupiter/*.0
Constructing namelist /var/crash/jupiter/unix.0
Constructing corefile /var/crash/jupiter/vmcore.0
100% done: 5339 of 5339 pages saved
.
.
.

How to Boot a System With the Kernel Debugger (`kmdb`) Enabled

This procedure shows how to load the kernel debugger (kmdb).

Note - Use the reboot command and the halt command with the -d option if you do not have time to debug the system interactively. Running the halt command with the -d option requires a manual reboot of the system afterward. However, if you use the reboot command, the system boots automatically. See the reboot(1M) for more information.

Halt the system, causing it to display the ok prompt.
To halt the system cleanly, use the halt command.
Type either boot kmdb or boot -k to request the loading of the kernel debugger. Press return.
Access the kernel debugger.
The method used to enter the debugger depends on the type of console that is used to access the system:
- If you are using a locally attached keyboard, press Stop-A or L1–A, depending on the type of keyboard.
- If you are using a serial console, send a break by using the method that is appropriate for your type of serial console.
A welcome message is displayed when you enter the kernel debugger for the first time.
```
Rebooting with command: kadb
Boot device: /iommu/sbus/espdma@4,800000/esp@4,8800000/sd@3,0
.
.
.
```

Example 9-4 SPARC: Booting a System With the Kernel Debugger (kmdb) Enabled

ok boot kmdb
Resetting...

Executing last command: boot kmdb -d
Boot device: /pci@1f,0/ide@d/disk@0,0:a File and args: kmdb -d
Loading kmdb...

Skip Navigation Links
Exit Print View
	Booting and Shutting Down Oracle Solaris on SPARC Platforms Oracle Solaris 11 Information Library