The callgrind output can be opened using kcachegrind. See the attd png for what I observe here. For some reason, inode_path is consuming unusual amount of CPU. This could be a resurfacing of a bug we thought was fixed by an earlier patch. Do you still have the logs for this particular run? Those are what we need to start looking deeper. Thanks -Shehjar Dave Hall wrote: > Hi again, > >> On Wed, 2010-01-06 at 14:38 +0530, Shehjar Tikoo wrote: >>> You can do this by using the following command to start up the >>> glusterfs client process: >>> >>> $ valgrind --tool=callgrind <path-to-glusterfs> -f >>> <path-to-client-volfile> -L NONE --no-daemon >>> <client-mount-point> > > Done. See attached. I must admit that I don't have much experience > with valgrind. I let it run for a few hours while I did something > else. When it was running the valgrind process was maxing a CPU and > cd /path/to/mount and other operations hung. I started it with > find -type f > /dev/null > > Cheers > > Dave
