La commande 'sysdumpstart' permet de provoquer un arrêt système avec génération d'un DUMP.
ATTENTION, cette commande arrête donc le système actif.
La syntaxe de base pour lancer le DUMP, sous le compte 'root' est la suivante :
# sysdumpstart -p
Une fois la partition relancée, il faut récupérer les informations pour un Snap:
# snap -r
# snap -ac
# cd /tmp/ibmsupt
# mv snap.pax.Z XXXXX.660.706.snap.pax.Z
Ci-dessous la documentation officielle en V6.1
sysdumpstart Command
Purpose
Provides a command line interface to start a kernel dump to the primary or secondary dump device.
Syntax
sysdumpstart [ -p ] [ -t traditional | -f { disallow | require_kernel | require_full }]
sysdumpstart [ -s ] [ -t traditional ]
Description
The sysdumpstart command provides a command line interface to start a kernel dump to the primary or secondary dump device. When the dump completes, the system halts. Use the kdb command to examine a kernel dump. Use the sysdumpdev command to reassign the dump device.
During a kernel dump, the following values can be displayed on the three-digit terminal display as follows:
|
Item |
Description |
|
0c0 |
Indicates that the dump completed successfully. |
|
0c1 |
Indicates that an I/O occurred during the dump. |
|
0c2 |
Indicates that the dump is in progress. |
|
0c4 |
Indicates that the dump is too small. |
|
0c5 |
Indicates a dump internal error . |
|
0c8 |
Indicates that the dump was disabled. In this case, no dump device was designated in the system configuration object for dump devices. The sysdumpstart command halts, and the system continues running. |
|
0c9 |
Indicates that a dump is in progress. |
|
0ca |
Indicates that a firmware-assisted system dump is not finished yet. System startup resumes after the dump completes. |
|
0cb |
Indicates that a dump is in progress. |
|
0cc |
Indicates that the system switched to the secondary dump device after attempting a dump to the primary device. |
You can use the Web-based System Manager Devices application (wsm devices fast path) to run this command. You could also use the System Management Interface Tool (SMIT) smit sysdumpstart fast path to run this command.
You can specify the -t traditional flag that allows to force a traditional system dump when the firmware-assisted system dump is configured.
Restriction:
- If traditional system dump is the current configuration, the sysdumpstart command cannot start a firmware-assisted system dump.
- If firmware-assisted system dump is the current configuration with an iSCSI software initiator dump device, the sysdumpstart command cannot start a traditional system dump.
You can specify the -f flag that allows to override the current full memory dump configuration.
Flags
|
Item |
Description |
|
-f{ disallow | require_kernel | require_full} |
Specifies if neither the kernel memory dump nor the full memory dump is allowed. If allowed, this flag specifies where the kernel memory dump or full memory dump is required. The -f flag has the following keywords:
|
|
-p |
Initiates a system dump and writes the results to the primary dump device. |
|
-s |
Initiates a system dump and writes the results to the secondary dump device. |
|
-t traditional |
Forces a traditional system dump independently to the current configuration. |
Security
Access Control: Only the root user can run this command.
Examples
- To start a kernel dump to the primary dump device, enter the following command:
sysdumpstart -p
- To start a kernel dump to the secondary dump device, enter the following command:
sysdumpstart –s
