[Gluster-users] unkillable processes

[Ian Duggan]

Hi,

I'm trying to setup gluster for use as a grid file system. I have 8 
machines, each serving one databrick. Machines 1 and 2 are serving 
namespace bricks, (afr'd to eachother). I run unify over the 8 data bricks.

Things are mostly ok, but I've been having issues w/ programs getting 
into unkillable states (I'm assuming TASK_UNINTERRUPTIBLE or something 
similar). It has happened to "find" and "du" at various points.

Is there anything I can do, short of rebooting, to kill/remove those 
programs? Restarting servers, clients, or mounts doesn't seem to help 
(esp since the unkillable file makes the mount unmountable).

--Ian