Panic strings


We've learned what type of system we are dealing with, and we know when it was booted and when it crashed. The next thing we need to know is how it crashed. Was it a forced crash dump due to a hung system or did the system panic or trap on its own?

Whenever a system panics, the actual panic message is kept in a string pointed to by a kernel variable called panicstr . panicstr will contain the starting address for the panic string.

Both Solaris 1 and Solaris 2 use panicstr for this purpose, so we will show an example from one OS only. As you can tell from the name of the unix file, vmunix.5 , we are looking at a Solaris 1 postmortem file.

Figure 9-5 Displaying the panic string
 Hiya...  adb -k vmunix.5 vmcore.5  physmem 15f90  *panicstr/s  _inprom+0x872: zero  $q  Hiya... 

In this example, we see that the system panic'ed due to what is called a "panic: zero." While it is possible for a system to actually panic this way, usually this is caused by someone using L1-A to force a hung system to panic and then asking the system to resume execution at location zero. This is an illegal operation and results in the panic shown above. So, until we actually look deeper into the system crash dump files, we will assume for now that the panic was forced by the system administrator in response to a hung system.



PANIC. UNIX System Crash Dump Analysis Handbook
PANIC! UNIX System Crash Dump Analysis Handbook (Bk/CD-ROM)
ISBN: 0131493868
EAN: 2147483647
Year: 1994
Pages: 289
Authors: Chris Drake

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net