[Linux-cluster] GFS: more simple performance numbers

[David Teigland]

On Thu, Oct 21, 2004 at 08:06:01PM +0800, David Teigland wrote:

> The dlm can get into a real problem if gfs hold "too many" locks.  If a
> gfs node fails, it's likely that some of the locks the dlm mastered on
> that node need to be remastered on remaining nodes.  Those remaining nodes
> may not have enough memory to remaster all the locks -- the dlm recovery
> process eats up all the memory and hangs.

In practice, I've actually run out of memory when rebuilding the resource
directory.  This could be helped by adding VMS-like "lock directory
weights".  And it can be a problem when nodes unmount (and leave the
lockspace); it doesn't require nodes failing.

-- 
Dave Teigland  <teigland at redhat.com>