On 08/18/2014 12:43 PM, Jaden Liang wrote:
Hello Gluster devel team,
I am trying to write a dummy xlator to learn about the developing
environment of gluster, and test this dummy xlator by reading/writing
some files via gluster.fuse. However, I found out that if there is some
buggy codes in my dummy xlator which will cause the glusterfs daemon
crash, the directory can not be remouonted via mount.glusterfs. Unless I
kill all the testing processes referenced the files or directories in
volume. How can I restore the mount daemon not by killing those
referenced processes?
Hi Jaden - welcome to gluster development!
Do your test processes change their working directory to the glusterfs
mount point? If yes, you can possibly try to refer the glusterfs mount
without changing the working directory of your test processes. With that
you should be able to umount a glusterfs mount point in case of a
glusterfs process crash.
-Vijay
_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxxx
http://supercolony.gluster.org/mailman/listinfo/gluster-devel
