Hi All,

Here is my situation:

I have standard 'a.out type' running processes that I would like to be able to suspend, restart my machine, and start up the processes again. The running process is long (20+ days) but takes up very little memory.

Is there a way to dump the job memory to hard drive and then after reboot load the memory again and restart the process?

I know I could probably do this using a virtual machine. But that's not the most convenient/elegant solution. Any suggestions?

Thanks.

What you want to do is called checkpointing or checkpoint/restart. Often checkpointing is done solely by the application program. Sometimes middleware and/or the operating system provides checkpoint/restart function calls to do some of the checkpoint/restart work for an application. I have worked with operating systems in the past which had checkpoint/restart. I also have worked with perpetually running middleware which could be restarted at a specified checkpoint when recovering from a crash.

Here is a list of checkpoint/restart software for Linux. They are a mixture of kernel based and application based checkpoint/restart software.

http://www.checkpointing.org/

-------------------
Steve Stites