Time-critical problem at Sun: exploding smbd memory usage

[David Collier-Brown]

Tony Shepard wrote:
| I applied the source diffs that Richard Bollinger sent around and
| recompiled 2.0.10 and it made a vast difference on the size of the
| daemon.
|
| pre patch:
|   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU
PROCESS/NLWP
|  8613 root       17M 1464K sleep   58    0   0:00.00 0.0% smbd/1

| post patch:
|   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU
PROCESS/NLWP
|  9097 root     3568K 1752K sleep   21    0   0:00.00 0.0% smbd/1

| I had dome some earlier testing and found that before applying the
| patch, the size of the daemon was determined by the number of
shares.

	Thanks, team: this is much more sane than before!

--dave
-- 
David Collier-Brown,           | Always do right. This will gratify 
Performance & Engineering Team | some people and astonish the rest.
Americas Customer Engineering  |                      -- Mark Twain
(905) 415-2849                 | davecb at canada.sun.com