I have some process (test) which runs on my machine. Sometimes it crashes and a core dump is generated.

Whilst the core dump is being written, the machine is completely frozen (even the mouse pointer cannot be moved).

I thought I could limit the CPU usage for the core-dump generation (I'm willing to accept a longer time to write the dump).

How can this be done? Is there a better solution?

I'm running Ubuntu 14.04.

  • of course, if you dont want the core, use ulimit -c 0 to suppress it. Or perhaps if you run your program under gdb you can exploit the error without causing a dump? – meuh Apr 6 '16 at 9:35

According to relevant bug reports (e.g., FS#41728 - [systemd] coredumps, 100% cpu usage, X hanging and systemd-coredump 100% CPU usage), core dumps in Linux are initiated in the kernel, with it copying the data out to userspace. That is where the time goes. You can generally interrupt things running as yourself, but not the kernel processes.

Further reading:

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.