HP 9000 V2500 SCA, Operator'S Manual

Page: 169 / 206

Chapter 7 147
Recovering from failures
Abnormal system shutdowns
Abnormal system shutdowns
Abnormal systems shutdowns (often referred to as system crashes) can
occur for many reasons. In some cases, the cause of the crash can be
easily determined. In some extreme cases, however, it may be necessary
to analyze a snapshot (called a core dump or simply dump) of the
computer’s memory in order to determine the cause of the crash. This
may require the services of the Hewlett-Packard Response Center.
V-Class servers using HP-UX Release 11.0 or greater employ a more
efficient dump mechanism than other HP servers using previous releases
of HP-UX. This mechanism is called fast dump.
Fast dump
When a system crashes, the operator can now choose whether or not to
dump, and if so, whether the dump should contain the relevant subset of
memory or all memory (without operator interaction).
By default fast dump selectively dumps only the parts of memory that
are expected to be useful in debugging. It improves system availability in
terms of both the time and space needed to dump and analyze a large
memory system.
The following commands allow the operator to configure, save, and
manipulate the fast core dump:
•
crashconf
—Configures the destination and contents of a crash
dump without rebooting. See the crashconf(1M) man page for more
information.
•
savecrash
—Runs at boot time and saves any information that may
be overwritten by normal system activity. See the savecrash(1M) man
page for more information.
•
crashutil
—Saves or manipulates the crash dump (if desired). It can
format the dump snapshot so that it can be read by the older
commands. See the crashutil(1M) man page for more information.
Installations that used to call
savecore
in any way other than by the
HP-supplied, unmodified
/sbin/init.d/savecore
script need to be
updated to use
savecrash
and/or
crashutil
.