methods.

Change a bunch of queries that join on the users table to use the idx
instead of the uid. This will need to be completed before we can really
archive users away, but close at this point.

current user has been loaded and checked. A little bit of refactoring
work as well.

only CatOS support)

node_type_attributes can now include a row with attrkey=use_acpi which can
either be set to:

  '1', if the FreeBSD image absolutely must use ACPI on this node_type
       in order to boot (e.g., our Dell SC440 machine)

  '0', if the FreeBSD image absolutely must NOT use ACPI on the node_type
       in order to boot (e.g., our pc2000s, possibly due to the IXP cards).

Otherwise, just don't set this attribute.

This value is queried by rc.frisbee and passed to slicefix after the disk
is loaded so that it can change /boot/loader.conf to enable/disable ACPI.

Note that ACPI (or lack thereof) has never been an issue with Linux or
Windows, so this is only meaningful to FreeBSD >= 5.0.

Note also that this does nothing for MFSes.  On machines that must use ACPI,
you need to use a special MFS for that node type (set via 'adminmfs_osid'
and 'diskloadmfs_osid' node_type_attributes).

 * support FreeBSD 6
 * client-side changes to support enable/disable of ACPI via slicefix
 * use dynamically linked Emulab binaries in frisbee MFS (for size)

most of the rest of the tables in the system (still a few exceptions).
Bound to be some bugs ...

This caused the reporting_uid field of 20,000 rows to be set to zero
instead of a string.

the logging. This cleans up a lot of code, and lets me (mostly)
encapsulate information about which log messages contain what
information in what version in a single place.

Database queries are handled by the library "libwanetmondb.pm", which
automatically gets the data from the ops DB or the datapository DB as
appropriate.

Queries to the datapository take a LONG time. The following
run took over an hour. The dstsite_idx (or srcsite_idx?) should be
added to the Datapository to make this faster. The ops DB is significantly
faster with this change.
  perl goodnodes.pl -e tbres/pelabbgmon -0 1167609600 -1 1167639600 5

leader, when changing the leader of a project when its created.

a bug, but I can't reproduce the old problem.

that we can run project approval from the command line. Part of the
ongoing push to get stuff out of php and into the backend ...

The command line is now:

	mkproj [-s] [-h leader_uid] [-m <message> | -f <file>] <pid>
	switches and arguments:
	  -s         - silent; do not send approval email to leader
  	  -h <uid>   - switch project leader to specified uid
	  -m <text>  - Include text in approval email message
	  -f <file>  - Include text from file in approval email message
	  <pid>      - project to approve.

Notes:

* The leader can be switched to a new user only at initial project creation.
  Once a project is actually approved (created), its too late. We need
  more stuff in place to change the leader after that, and that code
  is not written yet.

* Email is now sent from the backend script, so easier to recover from
  problems. When invoked from the web interface, the message text will
  be appended to the tberror email if the backend fails for some
  reason.  This should avoid the problem of that text getting lost and
  not being able to recover it.

* The web interface still handles part of project denial internally.
  Move that later.

Location of datastore is not "exp/datastore" not "datastore"
in exparchive.  Update Template::Instance::CopyDataStore to reflect this.

message.

* The unix group and gid are actuallt part of the group structure, not
  the project. So, when we load a project, load its default group as
  well, and have the routines for the name/gid hand off to the group.

* Change Lookup() to allow either a new idx or an old pid by looking
  at the format of the token. This will make it easier to transition
  to the new order of things.

* Part of the uncommitted uid changes is the addition of head_idx to
  replace head_uid. This code change is okay to commit now, since for
  the time being idx will shadow uid in all tables.

Changed the throughput calculation, cleaned up MaxDelay and
all other files in general to be more readable.

Added checks for out of order ACKs, and error conditions to
check for invalid packets.

sendto() calls block on Linux - verified this on
planet lab - made the writes to UDP socket non-blocking.

compliant with the Version 3 line format, as described in
libnetmon-format3.txt .

Note that I have made a couple small changes to the output format:
Write was renamed Send to be consistent, and I'm allowing Connected
events for UDP sockets, in the case that the application actually does
connect() the UDP socket. I'm not sure what we'd use this information
for, but it seems useful and is easy to ignore.

libnetmon is getting big enough that I should probably spend some time
soon doing code hygiene.