GitLab

pnode. Used to no be allowed, but with multiple routing tables this is
now possible. My code to construct link names on the fly when delay
nodes are inserted, was generating non-unique names which was causing
vlans to be dropped.

vnode and a PC node, under normal circumstances, a real physical link
is used. However, a user can use tb-set-multiplexed on the link that
allows assign to map multiple vlinks over the same physical link.
This is fine when the PC is running FBSD-STD which can use veth
interfaces. Not so if it is running Linux or a non-standard BSD.
This fix forces EMULATED links and VETH devices when the PC is running
FBSD-STD and turns those attributes off otherwise so that assign
will not give a mapping that cannot be realized on the client side.
For the gory details of this problem, look at testbed-ops email thread
"pcvm route problem" dated 29th Jan 2004.

A problem still exists in the non-BSD router + jail vnodes in a delayed
LAN. VLANs table is not getting the correct entries. A fix will come
soon. Follow the testbed-ops email thread "fix for the pcvm route problem"
dated 29th Jan 2004 for the gory details.

arguments, passes keyfile name, pidfile name, logfile name, etc.

Acts as a transparent filter for assign - parses the top file given to
assign, coarsens the graph in it, then feeds that to assign.
Translates assign's output back into context of the original graph,
and spits that out for parsing.

Right now, only includes a single coarsening algorithm - conglomerates
the nodes in leaf lans. Hopefully, it will soon include other
algorithms, such as running METIS to partition the graph.

Still needs a fair amount of work to be general - only works on
certain types of topologies right now.

Not called yet, but that will be trivial - just need to change
assign_wrapper to call assign_prepass instead of assign.

tmcd passes back an encapsulation flag to inform this.
NOTE: new version (15) for tmcd!

as well as links.

I am not using paperbag to run it on ops, but just installing the two
programs that are needed. Eventually should paperbag it.

This same script runs on both boss/ops and on experimental nodes to
fire off the linktest daemon on the nodes and wait. This is icky as
some crufty stuff has to be done so that it will run in both
environments. ltevent is of course not needed on boss, but do not want
to bother it now since it *is* needed on the nodes.

To run it on boss script needed severe cleanup and taint checking.
Also added the use of event keyfile, which is currently optional, but
needs to be mandatory once the images are updated.

The current problem is that run_linktest can hang waiting for the
clients to finish?. We can use quick mode in the experiment setup
path, but maybe we need a cancel operation?

into table_regex.

each veth from tmcd cause that is easy. Defaults to 1 for now.
Add tb-set-encapsulate to front end so Mike can play with it.

chance to commit it. Shashi, you should actually diff your changes
after a mysqldump to avoid picking up other stuff.

vnode weights (like CPU) and do a swapmod. This is a hammer but will do
for now.

local features/desires - turns out, we are using them in a manner
similar to types, so can save ourselves from looking at a lot of
bad solutions by being careful about going too far over on local
features.

Do some scoring, not just violations, for stateful features and
desires - this does a better job of nudging assign towards good
solutions.

The type each pnode is being used as.

The remaining capacity for all 'local' features.

Disallow _ in pid and eid since its illegal for DNS. Note, still
allowing it node names though.

for the normal case.

a type.

contents.  Make understandable: s/paper/boss/ and s/plastic/ops/.

Last-minute tweaks to error ranges (wider latency tolerance due to occasional errors when running at low bandwidth, plus wider Pathrate tolerance to avoid false positives while waiting for tuning data).
Added more info when aborting run_linktest.pl
Added more time to prevent a race condition between collector startup and test during the loss and bandwidth tests.

treated specially by the parser. Changed it to another bogus
hardwired IP!

Removed hardcoded "boss" from linktest.pl and run_linktest.pl
Changed historical log location.