I haven't removed the sshdport column from the nodes table yet, in
case I need to go back quickly.

Add some missing code to the check dbslot code to set the error strings.

nor having the source code in the proper subdir.

running on the node to force it to reboot.  Since planetlab has a bit of a
history of bugs/oddities in the initscript process (and thus slivers don't
always boot our stuff correctly), this commit special-cases the plab case
so that even if there's not an existing vnodesetup, we kill all the sliver
processes and rerun vnodesetup.

the files(dc.cc,dc.hh) were sent to me by Prof. Min Sik Kim( I attributed them to him in
the header comments ). Those files are used for wavelet-denoising and cross-correlation
calculation.

we have been making manually.

(i.e., the script will just run ntpdate)

tweak the disk afterward.

setting up.  First chunk: we don't want to skip straight to reboot if the
node is plab.  Second chunk: we need to run both plabnode alloc and remote
vnodesetup on plab nodes.  Third chunk: we have to ssh to the vnode, not
the pnode, on plab nodes, to make sure we make it to the correct sliver.
Not sure how this last affects jails, so I special-cased it for plab.

or BSD in partition 2.

stuff starting/done"), and another around the normal record commit
(using a zip file) ("commit starting/done").

requirements I started to make some notes on how the delay agent works
and how delays are setup.  These are those partial notes...

Removed leading "/" in:
   $page = "/template_show.php";
Since it broke things in the devel tree, in particular when clicking
on the GUID link under "Current Templates" in showuser.php3 goes to
  https://www.emulab.net/template_show.php?...
rather than
  https://www.emulab.net/dev/kevina/template_show.php?...

Currently we declare the valid types for an image to be those that are of
class "pc" and for which we have at least one node in the testbed.

With this change, when an image is created as a snapshot from an existing
node, we look at the partition info for the existing image to further
restrict the types that the derived image will work on.

This will prevent images derived from old images like RHL90 or FBSD410
from getting newer node types such as pc745d as a valid type.

One problem with the current solution is that, we don't know til the form
is submitted once, what node you are snapshot'ing, so the first time a
user submits the form, it is almost certain to come back with an error.

Further, they have no choice but to uncheck the offending types.  If part
of their customization was to add support for a new node type...too bad,
ya' can't add it!  There are numerous ways to fix this, all of which exceed
my Friday afternoon mad hacking skilz.