GitLab

Also added code to push the subtypes in the perm tables so I can stop
doing it by hand. Done badly though; needs cleanup.

sure they're within some tolerance of each other. Use this so that
cumulative float point errors don't cause us to spit out scary
looking false positive warnings.

Add plabnodes to stats gathering.

circumstances with vnodes, but I managed to run into it.

and push load data into the node_features table. Once every 5 minutes
from cron on boss.

turn on the 'unique' flag for widearea nodes.

but it's per node, not per type. Will initially be used to store
inforomation about planetlab nodes.

some reason, which is invisible to assign, that a node should never
be considered the same as another. This is for scoring, so that assign
doesn't prefer to give you 2 DSL nodes if you ask for 1 DSL and one
generic wide-area.

other, if both the feature and matching desire are found, their scores
are added together.

Additive feature names start with a '+'.

a pointer deref, but, thanks to the wonders of C++ operator
overloading, is actually a fairly slow function call.

nodes. The intent is to better support last mile types, which right
now are a total mess cause the types are associated with the virtual
nodes, and a node can have just a single type. This info has now been
moved in the node_types_auxtypes table and the node_types table.

Anyway, we no longer use wanassign on unconnected widearea nodes, but
use assign directly, much like we use assign to allocate virtual nodes
on local physical nodes. ptopgen inserts the physical widearea nodes
(even though they are allocated), and the proper counts for the types
that are available on them. assign will pick the nodes, and
assign_wrapper will allocate the necessary vnodes on the pnodes (but
in the case of wideare nodes, the underlying physical node does not
need to be allocated).

Also added some fixes for dealing with vtypes when used in conjunction
with widearea nodes. Rather then generating an error like it used to,
you can create a vtype in your NS file:

	tb-make-soft-vtype mytype {pcvroninet pcvwainet}
	tb-set-hardware $v0 mytype

and assign_wrapper now looks at the underlying types to figure out
what it needs. Note: No consistency checking yet; mixing a remote/virt
and a local/real type will break.

the plab vnodes according to link type (they are all of generic "pcplab"
type and class now), but add the linktype to the node_auxtypes field.  This
data in this table is then used to tell assign what kinds of types/roles the
node can take on.

preceeds the NS file. These are the auxtypes that a node in your ns
file is allowed to be, and comes from node_types_auxtypes. This is not
to be confused with node_auxtypes, which indicate which physical nodes
supply the desired types.

plab experiments.

Must be an integer between 1 and 1000. Okay, just kidding: 1<=x<=5

of all widearea nodes are now pcvwa/pcvwa, pcvron,pcvron, etc. The
last mile type (pcvroninet, pcwainet) comes from the aux types for the
node.

says to include widearea physnodes, with the types set to those listed
in the node_auxtypes table. Those types will include the last mile
types that a vnode on a pnode can take on. The :count value for the
types is set to the min of the number of *free* vnodes on the pnode,
and the virtnode_capacity in the node_types table. Obviously, if all
the vnodes are allocated, the physnode will not appear in the ptop
file. For plab, virtnode_capacity will get set to one in the
node_types table, thereby preventing "revisitation" of a plab physnode
in an experiment.

swapping.

This is primarily for the "queue an interactive job" form of batching,
but applies to all batch jobs at the moment.

This is primarily for the "queue an interactive job" form of batching,
but applies to all batch jobs at the moment.

dslice re-deployment.

creating the log directory for mysql, which needs to have the correct
permissions.

the types table yesterday, and it is hardwired here. Yuck.

naming scheme that ends in -A, where A is a letter, not a number.

footprint as it runs. All the credit for this one goes to Tim!

node_types table. This should make it easier to assign last mile
types to widearea nodes since a given type can now have as many
subtypes as you like.

Also add cpu_usage and mem_usage to the experiments table. The
user will be able to specify these in the NS file. Currently a
tinyint, where the default is zero, which means ignore.

to avoid hup-me-and-i-will-screw-ya mountd/NFS bug.

Use new -t option (use tmcd) in jails/plab

add -c option to force copying file over locally before untaring,
the copy code knows how to retry read operations that fail

stolen from Leigh's 'ixptest' experiment.

to do the fake links that were our old workaround.

that should be associated with a node when it is put into the ptop
file.

pnode's types were a subset of another's.

verification steps.