Collecting dump data

A dump might be critical for fault isolation when the built-in First Failure Data Capture (FFDC) mechanisms are not capturing sufficient fault data. Even when a fault is identified, dump data can provide additional information that is useful in problem determination.

All hardware state information is part of the dump if a hardware checkstop occurs. When a checkstop occurs, the service processor attempts to dump data that is necessary to analyze the error from appropriate parts of the system.

Note: If you power off the blade through the management module while the service processor is performing a dump, platform dump data is lost.

You might be asked to retrieve a dump to send it to IBM® Support for analysis. The location of the dump data varies by operating system.

For an AIX® operating system, collect the dump from the /var/adm/platform directory.
For a Linux operating system, collect the dump from the /var/log/dump directory.
For a PS703 or PS704 blade server that is managed by the Integrated Virtualization Manager (IVM), collect the dump using the Manage Dumps task in the IVM console.
To collect a system dump by using the Hardware Management Console (HMC), complete these steps:
1. Perform a controlled shutdown of all partitions.
  Note: A system dump will abnormally terminate any running partitions.
2. In the navigation area, open Systems Management.
3. Select the server and open it.
4. Select Serviceability > Manage Dumps > Action > Initiate System Dump. The dump is automatically saved to the HMC. For details on how to copy, report, or delete a dump after you have completed a dump, see Managing dumps.

Notices | Terms of use | Privacy | Support