GitLab

nodes that get moved to the reload pending experiment.  Without it,
tmcd won't free them from the reloading experiment.

Second, fix the case of freeing from the reloading or reloadpending
experiments.  The old code would get nodes stuck in their holding
experiment.

daemon take care of it.

- Changed internal representation to accommodate >1 vnodes per pnode.

- Added '-m' option to allow the user to specify
the maximum number of vnodes mappable onto each pnode.
(if -m is not specified, only one vnode is allowed per pnode)

- Added ability to "fix" a permanent mapping from a vnode to a pnode.
When specifying a vnode name, use the name of a pnode, with a "*"
prepended. This will be useful to fix one node as "The Testbed Node."

See comments at beginning of source for full documentation.

last_reservation, and letting the reload daemon pick it up from there,
we schedule a reload and stick it directly into the reloadpending
experiment. The effect of this is that the node never ends up in
the free pool before getting reloaded.

(Note: We _don't_ do this if the node is being freed from the reloading
or reloadpeding experiment)

This change may not be permanent.

(/etc/rc.d/testbed/slothd rather than /etc/testbed/slothd)

For consistency's sake.

Slothd changes to force an initial wait of <interval> (def 1hr).  Selectable
with -w.  Also changed to scan ALL ttys by default (can use UTMP with -a).

Finally, deploy script changed a bit to be more robust wrt down nodes.

Jay's request. Its all I can handle before coffee.

more daemons and tcp connections ...

(the base experiment). Also do the proper calculation when inserting
into virt_lans for delay and plr, which is different then when
updating the delays table directly. Change from using ssh to using the
event system to make the change on the node.

were introduced.

Look at comment at beginning of wanlinksolve.cc for full documentation.

Also added -v switch to turn on verbosity (now off by default.)

too much CPU

of just pinging it, poll its eventstate in the database, and wait
until it's reported ISUP. This way, we don't report that a node is
ready before it really is. Also bumped up the timeout by a couple of
minutes to account for the extra time it takes for the OS to boot.

Right now, if the OS is considered pingable, we assume that it will
report ISUP. In the future, however, when the node state is more
formalized, we will have better ways of determining this.

the warning and update the info we are given, the other fields will use
the default until otherwise set.

Reset the logfile on a RESET event.

ahead and call it as before but we don't exit if teardown fails.

Change max packet size from 8k to 64k.

DON'T FORGET TO REMOVE YOUR OLD tg2.0 SOURCE TREE SO THE PATCH GETS REAPPLIED!

stsyle routes for nodes that have multiple interfaces (thats a
restriction in our parser borne of expediency). So, when routing to
nodes that have more than one interface, force a net route instead of
a host route. No biggie.

better picture of where time is spent during experiment setup.

routes in NS, BSD, and Linux formats. Use with -t to toggle the
network optimization.

Is controlled by a new configure variable called TIMESTAMPS - if this
variable is set, prints out a timestamp, along with its arguments (so
that you can distinguish between the timesteamps.) Otherwise, does
nothing. The basic idea is that we can sprinkle calls to this in the
code, and only enable it on devel trees.

(we should remove the old ones soon)

These allow zooming and present (optionally)
  - Node type
  - Node IP listing
  - Link characteristics
     - always shows bw
     - shows latency if > 0
     - shows loss if > 0%

Added links in shownsfile.php3 for specifying zoom and more/less detail.

Added new vis tools.
(we should remove the old ones soon)

parser runs. The staticroutes script is a wrapper for Chad's route
solver. The network optimization is currently turned off; use -t to
turn it on, until I know if its correct.

Note that Chad gets credit for routecalc.cc; I'm just committing the
file for him, with a couple of trivial changes that I made.