experiment cause of failure!

get them from veth_interfaces, but not bothering.

direction of an emulated link.

pc3000 nodes using a special MFS.

use the new type with vnodes.

table for the appropriate pcvm types.

the tail file of course. Called from TBdbfork() in libdb, EventFork
resets the event handle so that the child does a reconnect. Note that
I do not disconnect in the child since I have no idea what that is
going to do to the parents connection to the elvind, as Elvin makes no
mention of what to do in the presence of a process that forks after
connecting to the event server. At the least, this avoids a bunch of
warnings and errors from vnodesetup!

<a href="http://www.ee.technion.ac.il/people/idish/Abstracts/Drum.html">
"Exposing and Eliminating Vulnerabilities to Denial of Service Attacks
in Secure Gossip-Based Multicast."</a> To appear in the
<cite>International Conference on Dependable Systems and Networks
(DSN 2004)</cite>, Florence, Italy, June-July 2004.

the link agent). You can now do this:

	link_config -s nodew1 testbed three-wireless lan0 ENABLE=yes

to bring a link up (or down; ENABLE=no). This just gets passed along
in the event arguments. Basically an alias for "tevc now lan0 UP/DOWN"
to make things a little easier. At present, wireless lans are brought
up/down with ifconfig since "txpower off" does not work properly.

I suppose I should add the equivalent changes to delay_config ...

was giving me multiple rows for nodes that have multiple interfaces in
use!

Transmogrified underlying data structure to a tree representation for flexibility reasons. Added a new Assigner module which calls the partitioner repeatedly to form a hierarchy.

Switch around examples at the end.

use a non-production channel.

1. tbbootconfig: ensure block is zeroed on first init, fix cut/paste error
2. rc.frisbee: cleanup interface to slicefix
3. slicefix: cleanup, make it work correctly, init tbboot block for cd boot

Unrelated:
1. link-agent makefile: build link-agent when doing client-install

floor.

* On the shownode page include the channels that a wireless interface
  is currently set to, by looking at the interface_settings table. So,
  if you click on an allocated node in the wireless map, you can see
  what channels it is using.

  Still to be done; map between a channel number and a frequency.

* On the floormap page, create a small table at the top of the page,
  listing what frequencies are in use on each floor. This is rather
  primitive of course, especially the part where I assume we have just
  one building, but it will do for now.

This is not complete. We probably need some sense of scale on the
maps.

Needed for new PCs with SATA disks which start as "ad4".

'loadinfo' request.

project permission. By default you get a summary of all the free nodes
you can use in all your project memberships, which is not too
useful. It would be more useful to see what types you can use by
project, since that is how permission is granted and how experiments
are created. So, when you are a member of more than one project,
display a list of those projects along the top so you can see the list
on a per-project basis.

More generally, if you are an admin person:

	www.emulab.net/nodecontrol_list.php3?showtype=summary&bypid=tbres

will show out the summary information for a particular project. At
some point I need to make this a subroutine that we can call from
other pages to drop in a summary table.

system as well as the summary page.

For tbres/resweb (75 nodes):

	Old time: 42 seconds.
	New time: 19 seconds.

For testbed/BigLan (600 nodes):

	Old time: 120 seconds.
	New time: 40  seconds.

So, still needs some work, but for the average case its not so bad
anymore.

I also turn on compression to level 5 by default in our sshxmlrpc.py
library, thereby overriding the user's config file for that option.

delays to get the wrong BW setting in the delays table.

You can look at the error code to figure out if it will ever be
possible to map this experiment onto the testbed.

log, and to look at the summary page when the problem is not enough
free nodes!

dequeued.

inactive experiments.