HP 9000 V2600 SCA - Manual (Page 174)

152 Chapter 7
Recovering from failures
Abnormal system shutdowns
/etc/rc.config.d/savecrash ) reduces system recovery time. After
the system recovery, run savecrash manually to copy the memory
image from the dump area to the HP-UX file system area.
Partial save
If a memory dump resides partially on dedicated dump devices and
partially on devices that are also used for paging, only those pages that
are endangered by paging activity can be saved.
Pages residing on the dedicated dump devices can remain there. It is
possible to analyze memory dumps directly from the dedicated dump
devices using a debugger that supports this feature. If, however, there is
a need to send the memory dump to someone else for analysis, move the
pages on the dedicated dump devices to the HP-UX file system area.
Then use a utility such as
tar to bundle them for shipment. To do that,
use the command /usr/sbin/crashutil instead of savecrash to
complete the copy.
Crash information integrity
This section discusses how to make sure the part of memory that
contains the instruction or piece of data that caused the crash is
captured. The factors that must be considered are:
• Full dump vs. selective dump
• Dump definitions built into the kernel vs. defined at runtime
• Using a device for both paging and as a dump device
Full dump vs. selective dump
The only way to guarantee capturing the specific instruction or data that
caused the crash is to capture everything. This means selecting a full
dump of memory.
Be aware, however, that this can be costly in terms of time and disk
space. A large amount of time and disk space is needed to dump the
entire contents of memory in a system with a large memory
configuration or to copy a large memory image to the HP-UX file system
area during the reboot process.
The amount of dump area should at least be equal to the amount of
memory in the system; depending on a number of factors, additional disk
space greater than the amount of physical memory in the system may be
needed