Upon completion of this module, you should be able to:
1,describe how a process creates a system dump
2,configure a system to collect and store core files
3,differentiate a system panic condition from a system hang
4,list the steps to perform an initial core dump analysis
5,use the adb and crash debuggers to analyze crash dumps
The following events occur in response to a system panic:
display a panic message at the console
perform a stack trace, listing routines that led to the panic
dump "interesting" portions memory to the swap device
reboot and copy the core file from swap to a file system
The severity of the hang needs to be monitored by trying diffrerent methods of system access, and varies depending on the cause of the hang, which commonly include:
an application has hang
a terminal or window has hang
the system has hang
the system is overloaded and has a resource bottleneck
The options available with dumpadm to change the crash dump and savecore configuration details are:
-c content_type
-d dump_device
-m min
-n suppress the automatic execution of savecore on reboot
-s savecore_dir
-u update the configuration details based on the contents of /etc/dumpadm.conf
The following command specifies collect the core file on swap, populate the file with kernel pages only, and save the core file to /opt/crash/system1 on reboot:
# dumpadm -c kernel -d swap -s /opt/crash/system1
To invoke adb for crash dump analysis,type
# cd crash_directory
# adb -k unix.n vmcore.n
To start adb on a live system, run the following command:
# adb -kw /dev/ksyms /dev/mem
阅读(3016) | 评论(0) | 转发(0) |