[Linux-cluster] GFS Locking questions?

[David Teigland]

On Fri, Oct 01, 2004 at 04:56:46PM -0600, Matthew B. Brookover wrote:

> Similar problems happen if you run readlock and writelock on the same
> host.

The code in cvs now correctly sets EAGAIN for SETLK.

There are more fixes and improvements for posix locks that require a 2.6.9
rc kernel.  I'm not sure to what extent they may help you.  These are the
changes for gfs, lock_harness, lock_dlm and lock_nolock if you'd like to
try.  The patch is against the current cvs:

http://people.redhat.com/~teigland/plock-2.6.9.patch

-- 
Dave Teigland  <teigland at redhat.com>