Queries the ops DB to make sure data is numerous and recent.

  different than tbres/pelabbgmon
- Added ability in manger.pl to kill a node's bgmon.

file. gcc 4.0 (used by FC4) didn't like having it defined in the .h
file.

Just wanted to make sure i checked in everything used to produce the final
flexlab udp results.

1) Packets due to EWOULDBLOCK are now being considered as losses.
2) Udp socket in KernelTcp is now non-blocing.

own flexlab stuff for the time being.

2) Removed some unnecessary debug statements.

1) Fix max delay event to include the forward path min delay
2) Fix time conversion problem to ulonglong in packet sensor.
3) Fix averaging sensor to better average in case of heavy packet loss.
4) made the buffer used to fill in junk udp packet data is a static member - takes up less memory - and allows up 65k sized udp segments to be sent.

socket 'names' backwards.

Also changed so to add a new 'Accepted' report type to distinguish it
from Connected.

printed twice.

Fixed bug which caused reading of uninitialized variable and thereby caused TCP connections to all be rendered invalid.

Applied Pramod's fix of the pcap bpf expression to TCP. Abstracted and removed duplicate code. This turns out to fix some puzzling behaviour which we didn't realize was related.

The monitor will now ignore IP addresses that are not within its ip-mapping. This includes localhost. Updated libnetmon-format3 to point out limits in keysize and virtual-connections/socket.

2) mindelay/maxdelay are now being sent in millisec. mindelay is
now being set on the reverse path too.
3) Added a check - not to send zero bandwidth values to the monitor.
4) Removed some debugging print statements from KernelTcp.

	run make-tarball to create a new tarball
	run install-tarball to get it out
	run start-experiment
No swapout/in should be necessary as long as you are already using magent
(udp server is started out of auto-magent.sh)