From: Richard Hu (rhu_at_opnet_dot_com)
Date: Thu Mar 03 2005 - 07:16:08 PST
If attempting to restart a process results in the following error: "cri_syscall(CR_OP_RSTRT_REAP): Invalid argument", then how can I attach a debugger to it? Thanks, Richard At 09:01 PM 3/2/2005, you wrote: >Richard, > > By its nature, cr_restart has to do some pretty "unconventional" > stuff. However, I am suprised that it is crashing. The program > cr_restart is really only a requester for the restart, so I am not sure > how valuable it is to run gdb against cr_restart. What you really > probably want it a way to attach the debugger to the restarted process. > We don't have any currelty known issues or work-arounds related to > debugger use. I encourage you to enter into bugzilla a detailed > description of your problem and we'll see if it is something we can address. > >-Paul > >Richard Hu wrote: > >>Hi, >> >>I had a question about debugging when attempting to restart a program >>using your Berkeley Lab Checkpoint/Restart program that I was hoping you >>could answer. I wrote a simple program that assigns a value to a >>variable, checkpoints the program and then assigns another value to the >>same variable. When I run the program, it checkpoints properly and I can >>use cr_restart to restart it correctly. However, when I attempt to step >>through cr_restart using a debugger (gdb in this case), it crashes. >> >>I searched through bugzilla for any relevant information and could not >>find anything relating to gdb. Are there any known issues involved with >>attempting to use gdb to step through cr_restart? If so, are there any >>work-arounds? >> >>Thank you very much >>Richard Hu >>rhu_at_opnet_dot_com >