boss.

regen the ssh key if its missing. If there is still an authorized_keys
file, cat the new identity.pub onto the end of the authorized_keys
file.

experiment table fields that hold the solver weights that are set by
the user in the front end with the tb-set-wasolver-weights command.

state to REBOOTING, and then wait for the ISUP state to be set.  This
change reflected in the clientside startup scripts on remote nodes,
that now issues a REBOOTED event, and then an ISUP event after
everything is setup properly.

conditionalized on isadmin() status. These pages were mostly
overlapping anyway, and the nodecontrol_list.php3 page was done more
cleanly. No point in all the duplication. Dropped the Node Control
menu item from the list.
Add explicit "widearea" view to get remote nodes in their own page,
instead of listed after all the PCs. Also some minor formatting
changes.

project list for each user, do one big query for all users at the
beginning of the script.
Also some minor format changes.

watchdog at the moment, but it will be. Right now it does boot time
stuff; issues tmcc state event so the testbed knows (REBOOTED), does
an account update to get any accounts missed while dead, then sets up
and vnodes (tunnels and such) that where supposed to be running on
the node, then issues a tmcc ISUP event.

After that, goes into a loop doing periodic account update. At some
point it would be good to look for stale vnodes (that could not be
torn down because of network connectivity problems), but there are
some race conditions that I need to work out first.

node reboots. Currently, starts the watchdog daemon.

instead of account update. Doing an account update at boot is nice,
but hardly sufficient.

instead of waiting. This option used by new watchdog daemon that
runs on the RON nodes. No point in waiting if another is already
happening.

more thought, so just leave the directories.

files are generated.

	tb-set-wasolver-weights delay bw plr

Where I am told proper defaults are 1, 10, 500. Setting any to zero
effectively drops that metric from consideration when the solver runs.
I've updated the tb_compat files and the documentation.

each link.

Now, the problem with running tbreport on experiments that have
wide-area links, but are not swapped in, should _really_ be fixed.

 - Make configurable options specifiable on the command line (idle
   time, packets/hour, "freshness" time).
 - Check how fresh our data is, and mark it as stale if it is old.
 - Add display options. Default is show only inactive, swappable, non-stale.
   Use -s flag to show stale expts too, and -u flag to show unswappable
   ones. (-s and -u display whether they're active or not)
 - Changes in the database optimized the queries for me, so instead of
   taking 10-15 seconds to run, it now takes 1.
See usage (idlecheck -h) for more info.
We're now ready to put this in a cron job and in idle view of the expt
listing web page.

experiments early enough.

Then, switched to using the new TBWideareaNodeID call so that we can
report links that go between RON nodes and the testbed properly.

Also, made a new message, seperate from "No data available" for vnodes
that are on the same physical node, or that seem to be at the same
site.

returns the node ID that should be used in the widearea_* tables.
This is mainly so that we don't have to hardcode 'boss' as the local
representative in too many places (all local node are assumed to
have the same network characteristics as boss.)

with tb-set-wasolver-weights.

Properly detect and report failures while trying to start slothd, and remove the lock file before trying to start it if it isn't running.

that batch system will not be confused.

Return an error if a negative weight is passed to -1, -2, or -3.
Also, interpret negative values of minrounds as zero.
Also, interpret any negative value for maxrounds as -1.

Added loss metric (-3 switch; must be used after -2 switch.)..
Also added -1 switch to specify latency weight.

Add indexes by mac to interfaces and iface_counters, and node_id index to the latter as well. Speeds up idlecheck from 10-15 seconds to 1-2 seconds.

currently assigned virtual nodes, say when a node reboots.

Also, changed the way that bandwidths were being looked up in
the widearea_delays, to make it work better if nodes' IP addresses
change.

- killing off any currently running slothd (potentially started by rc.testbed)
- removing the lockfile
- staring the daemon

Slothd should now startup properly after a machine crash.

period over which the data was collected.

reported round-trip, and NS link delays are one-way, it was hard to
figure out how well they matched. The column header reflects this. Not
totally realistic, but should make the report much easier to read.
Suggested by Tim.

Also, added a check to avoid priting widearea links if the experiment
is not swapped in.

Current settings:
 - 4 packets/hour avg over last 24 hours (100pkts) is activity
 - use of a tty in last 24 hours is activity.
 - Latest report from slothd must be within last 120 minutes.

Caveats:
 - Shows all expts that are not known to be active. This includes
   expts that we don't have any information on. I plan to fix this so that
   expts that are reporting are very different from those without reports.
 - Because of the above, we're quite conservative in marking nodes
   active. When the above gets fixed, this will be better.
 - Displays unswappability. We may in some cases still want to harass
   owners of unswappable expts, since they can mark it arbitrarily.
 - Currently takes ~10 seconds to run. I plan to run the queries
   through explain and try to optimize them a little better, or even
   simplify where possible.

The output of this script is what will eventually get used in idle
view of the web page. Unless optimization makes this script run in
less than 1 second, we'll most likely need some sort of caching of
these results before being able to use them in the web page. Perhaps
stashing them in the db with the time the idleness was detected, so it
is easy to get to.

Suggested by Tim.

links in an experiment, the actual measured delay, bandwidth, and loss
for those links will be displayed.

This fix is more complicated that the original changes that caused the
problem :|