UNIX for Dummies Questions & Answers

I have a process(c program) which runs as daemon and is causing memory leak.
Is there any way to detect memory leak in this already running process?
Just like we attach gdb to a live process to debug it, can I use valgrind to detect memory leak in such live process?

When I tried using valgrind to such a live process it gives me error "< Process Name> instance already running...".

How do you know the leak is happening?
Does your process crashes everytime it's run after certain time or is there a trigger which results into a crash/abort for "ENOMEM" ??

The output of top command for the process daemon shows an increased memory usage from 6 mb the normal usage to 800mb when the daemon is utilized.
After the daemon utilization completes still the memory usage shown is 800 mb and it does not drop down.

Ok. Yes seems to be memory leak case.
But then what is the trigger or is it always expanding in every run?

If you know the trigger which is causing the issue, you would easily zero in on to the affected code area.

So far I'm not aware if Valgrind could be attached to a running process like gdb but I know that it could generate good enough stats with function wise memory usage reports.

Need to look at the Valgrind manual.

Why, btw, you wanted to attach the daemon when you could have just run the same in the Valgrind environment from the command line itself?