Skip to main content

8.13.3 Points to Note on Using Hypervisor Dump


8.13.3 Points to Note on Using Hypervisor Dump
Note the following points on using the Hypervisor dump function.
  1. Even if an error occurs in Hypervisor, no dump file can be collected when the directory for saving dump files has:
    - Eight existing dump files, or

    - Insufficient disk space for the /var/opt/SUNWldm directory

    Routinely check the number of dump files and disk capacity of the directory for saving dump files to confirm that the requirements for dump file collection are met.
  2. If a dump file cannot be collected for a reason such as the above, processing stops. After solving the problem, execute the ldm start-hvdump command to collect a dump file. If you restart the control domain without executing the ldm start-hvdump command, no dump file is collected, and the logged Hypervisor information is lost. Even when the hvdump-reboot setting is on, if a dump file is not collected, the physical partition cannot be restarted. In such cases, move or delete a collected dump file first. Then, use the ldm start-hvdump command to collect a dump file. After the dump file collection completed, restart the physical partition.
  3. Even when the hvdump-reboot setting is on, if the OpenBoot PROM environment variable auto-boot? is false, processing stops at the ok prompt without restarting Oracle Solaris.
  4. If the control domain is unintentionally started in the factory-default configuration, Hypervisor dump processing may not complete. This may be because the control domain is started in a special factory-default configuration to collect a Hypervisor dump. Execute the ldm list-hvdump command to check whether a dump file has been collected.
    The following example shows that a dump file could not be collected.
primary# ldm list-hvdump
hvdump=on
hvdump-reboot=on
Pending hvdump exists