Service manual

Troubleshooting 3-77
Troubleshooting a hang is difficult. The suggestions in Table 3–12 are intended
to give you a start.
There are some causes you can eliminate. Theoretically, at the hardware level,
the system should not hang. Transactions are tracked such that if one is not
making forward progress, a timeout is triggered, a machine check is generated,
and the system crashes. Such an event is a fault and is considered a serious
systemwide event that causes PSMs in the system to initialize (except for error
state) and reset all components (ASICs and CPUs) in the system. All QBBs
reset. When the machine re-boots, the PALcode attempts to collect the error
state, if any, in control and status registers and build a system machine check
(660) error frame that Compaq Analyze will automatically decode. See Section
3.12 for information on running Compaq Analyze.
At the operating system level, there are timeouts in software that get triggered
that also cause crashes. Applications may hang but they can be handled at the
operating system level by stopping the application.
If the microprocessors on the CSB lock up, the system could be running but
access to it, through the console, may not function.