file: we return info for boss/ops in the "hostnames" command, but only if
the command would have otherwise returned something.

This is because the default hosts file we distribute in the rootball already
has boss/ops in it.  But if, during bootup, tmcd returns hostname info, that
hosts file gets overwritten and boss/ops info gets lost.

I could just grep out the info from the original hosts file and transfer it
to the one we are building, but what do I grep for: "boss", "ops",
"emulab.net"?  Don't really want to hardwire any of those in the client-side
script.

By returning this info via tmcd, I also don't have to modify the client-side
and thus don't need to build a new rootball!

annotation.

for Eric.

show pages. The annotations are stored as specially named metadata
items.  All this will change, but this was a quick hack. The UI needs
work, and we probably want to support multiple annotations later, but
this will do for now.

objects, not sockets.  The reason for doing this is because I expected to
be talking to the new PLC4-era NM over an ssh tunnel, which would only
allow me to talk to the NM over stdin/stdout (i.e., no ssh socket tunnels
because there's only a pseudo shell on the NM side that proxies
stdin/stdout to the NM).

I was trying not to have to reimplement the whole httplib, so I overrode
the bits that can't stand to use anything but a socket.  Lots of those.
Anyway, seems to work... so anytime you need to talk XMLRPC over HTTP but
not over a socket, this will work for you.  You'll just want to write your
own "real transport" method ... I only wrote one that popens a command.

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.

template.

On dark monitors (like Eric's), this will hopefully make the box a
little easier to read. This is about as bright as I can make it while
still looking vaguely blue-ish.

expired from people's caches by now.

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.

Also remove guid/version argument requirement for template_export.

do the right thing. No such luck, giving up for now ...

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)

time, even when just a single param.

Added non-blocking i/o to monitor. Takes care of 'improper header size' errors. Also, added protocol information to set_connection calls so that udp is recognized on delay nodes. Removed a str-to-int conversion bug.

UDP code should now be fully funtional. ( except that UdpServer is not yet
run automatically on plab nodes...)
2) Added an option to UdpServer to use min=3 redundant Acks by default and to send
as many as possible with -maxAcks command line option.

when the experiment is deleted. This has not been happening for
templates since the archive is shared.