HP-UX HB v13.00 Ch-08 - Crash Dumps

HP-UX Handbook – Rev 13.00 Page 5 (of 38)
Chapter 08 Crash Dumps
October 29, 2013
There are three different types of crash events: PANIC, TOC and HPMC:
PANIC
The crash event type panic refers to crashes initiated by the HP-UX operating system (software
crash event). We differentiate between direct and indirect panics.
A direct panic refers to a subsystem calling directly the panic() kernel routine upon detection of
an unrecoverable inconsistency, for example:
 panic ("wait_for_lock: Already own this lock!");
 panic ("m_free: freeing free mbuf");
 panic ("virtual_fault: on DBD_NONE page");
 panic ("kalloc: out of kernel virtual space");
An indirect panic refers to a crash event as a result of a trap interruption which could not be
handled by the operating system. For example when the kernel accesses a non-valid address
(virtual or physical), a Data page fault (trap type 15) would result. The trap handler will save
some state information (aka save_state) and then call the panic() routine to bring the box down in
an orderly manner. This is indirect since panic() is called at a point slightly later than the trap
condition that caused the failure. Some examples
 trap type 15, Data page fault
 trap type 18, Data memory protection fault
 trap type 6, Instruction page fault
TOC
The crash event type TOC refers to crashes initiated by a Transfer-Of-Control sequence. There
are three different ways of getting a TOC event for a CPU:
 Operator initiated TOC (e.g. manually pushing a TOC button, or cycling the
power button 3 times on some systems, or using the TC command in console
mode).
 MC/ServiceGuard initiated TOC (e.g. when it is unable to maintain contact with
the cluster daemon).
 Crash path initiated TOC. On multi-processors systems, the processor taking the
initial crash event (e.g. a panic) will cause the other processors to perform TOC
automatically.
A manual TOC is usually done when a system is hung or unresponsive. This way the
crash dump can be analyzed to determine the root cause.
HPMC
The HPMC crash event type refers to High Priority Machine Check crashes initiated by the
hardware due to hardware inconsistencies or malfunctions such as a Data Cache parity error.