Also rearrange configure.in to produce unattended answer files based on
which Windows keys are present in the defs file.

This will prevent us from hanging waiting for ops/fs filesystems until
we have at least started up critical services like named.

Make sure sitevars get initialized on initial installation of an Emulab.
Fixes to the update_sitevars script, mostly in case we someday want to
run it on every testbed software install (which we do not do right now).

For ops and fs there is a race with boss that prevents us from locking
down ARP entries early. For now, we do the lock down later in the boot.
If someone spoofs boss or the gateway before then, we will detect it
when we request the ARP info via SSL-enabled tmcc.

This officially drops the pretense that fs nodes can operate with minimal
Emulab software. If you have a seperate fs node, it had better be dedicated
to Emulab!

However, it still doesn't do everything. In particular, accounts are not
installed. This has never been needed for serving NFS, but is needed for
the samba stuff to work correctly.

Also, you cannot do an fs node software install from boss yet as we do not
mount fs filesystems on boss. You really cannot do a full ops install from
boss either since we don't mount ops' /usr/local/etc/emulab directory.

Erroneously skipped installing dijkstra on Windows.

The 'storagehost' value in the erole enum column will allow ptopgen to
identify hosts that are eligible for mapping as an infrastructure
storage host.

a XEN based ElabInElab with boss and ops running firewall rules based
on the ruleset we use on Utah's firewall. To turn this on, add this to
your NS file:

tb-set-elabinelab-attribute CONFIG_FIREWALL_BOSS "yep"
tb-set-elabinelab-attribute CONFIG_FIREWALL_OPS  "yep"

You do not have define both.

does and how it does it.

you run two at the same time. So we have to serialize the
calls. Rather then worry about each call, just take a big lock here.

The 'storagehost' value in the erole enum column will allow ptopgen to
identify hosts that are eligible for mapping as an infrastructure
storage host.

swapmod to leave behind vlans on the switch. As a side effect,
add a "class" field to the vlans table so that we can determine if a
stale vlan is a control or experimental vlan (since we now support
building experimental vlans on the control stack, so the stack no
longer tells us what kind of vlan it is).

so that we can do post-mortem debugging. I'm adding this so I can
track down the current snmpit failures, but might useful later too.

Sadly, mysql dump files have no table metadata so once the schema
changes, these files will become less useful if we ever want to load
them into a running DB. It would be nie to dump them in a format that
allows this, but will be a bunch more work.

containers.

This is a hack.  The "blockstore" procedure is just returning a node
object, which causes skew in the resulting topology vs. what is generated
by the implicitly created node objects in the real testbed parser.  Harmless,
but not a good validity test.  Need to revisit.

for mere users.

This makes it convenient to request public addresses for virtual nodes
through the NS interface, like this:

    tb-set-node-routable-ip $node 1