Info on Core dumps

If a program/process fails, and it has been designed to do so, it will dump the state of the process running in the machines memory. This is a core dump. Through core dump analysis, you can the then work out why/what happened to make the process fail.

I am not sure that you can use a core  dump to restart a process. You would normally use tools to analyse the dump. A typical Solaris OS dump can be 30meg plus, and is not something you can read normally.

A process can dump wherever it has been configured to dump, but for example, a traditional Solaris dump would be put in /var/crash/<hostname> and be called unix.N where N= the crash no.

You aren't going to find any method of restarting the crashed process using the core file that I can think of.. if you have the correct resources, you could use a debugger like adb to examine the core file - if you just want to look at which executable caused the core, issue "file <corefile>" - it should tell you some linking information, as well as the name of the executable that left the core.

core files are generally more useful to developers with the source of the program, so they can determine where and why the crash took place.

Hang on- you just copied what I said

Sorry Jaran, but Jonke got there first.
Jonke, if you'd post your comment as the answer please !

Thanks

Penfold

Sorry, was hoping to add a little depth, not take credit.

Just as a side-note...depending on the application, core files can be considerably larger than 30MB. They really are for professional analysis. Example on our systems using SAP R3 ..a typical core is about 300-800MB...not much chance of pulling up that file in an editor.

Think you have what you need from above ...good luck !

visit http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xc2565c7609e9d61190050090279cd0f9,00.html for an indepth info on analysing core files.

HTH