A key part of recovering from a hung system is collecting system crash dumps and core files.. follow the steps listed below:

How do I recover from a hung system?

Step 1:

Collecting system crash dumps and core files are important steps for diagnosing system crashes and application failures. In addition, forcing a system crash dump is sometimes needed to recover from a hung system.

The dumpadm command with no arguments is used to view the current dump configuration.

Step 2:

The dumpadm command above is used to change the dump device to the second disk drive, slice 6.

Step 3:

Note the output shows the dump device is successfully changed to the second disk drive, slice 6. All previously unwritten system buffers are then flushed out to disk with the sync command, ensuring that all file modifications up to that point will be saved.

Step 4:

The savecore  L command forces the kernel to save a live snapshot of the running system and write out a new set of crash dump files.

Step 5:

Note the dump succeeded in the output. Next, the files in the /var/crash/proto145 directory are listed to ensure the crash dump succeeded.

Step 6:

The output shows the saved crash dump files, vmcore.# and unix.# in the directory.

Step 7:

Note the default initial configuration output. Next, a corefile directory is then created, as well as a global core file path. Logging is also turned on to generate a message when a global core file is attempted.

Step 8:

The configuration information is displayed to verify the previous changes using the coreadm command.

Step 9:

The output shows that the core file directory and a global core file path have been created. Logging has also been turned on. Next, another terminal window is opened and a directory called /var/tmp/dir is created and made the current directory.

Step 10:

The pwd command is used to confirm that the current directory is /var/tmp/dir. Next, the ps command is used to find out the PID of the new shell.

Step 11:

The kill command is used to send a signal 8 to the new shell. Signal 8, or SIGFPE forces a new core file.

Step 12:

Note that the kill  8 command terminated the shell and the terminal window in which it was executed. In the original window, the file command is used to verify that a new core file exists in the current working directory.

Step 13:

Note the output displays information about the core file. The /var/core directory is then listed to check for a core file.

Step 14:

Note the output indicates that there is a core file within /var/core. Next, observe the messages generated in the messages file due to the fact that coreadm logging is enabled. The tail command is used on the /var/adm/messages file to observe these messages.

Step 15:

Note the logging output states information about the crash dump and core files.

End:

For more in-depth Web Based Training on variety of related topics, visit Sun Spectrum.

Solaris 10 Operating System

This solution has no attachment