GitLab

It is now only 1/3 slower than the original. The price of modularity. *sigh*

ptopgen_new is now merely twice as slow as the original.

image_history table.  Output a warning if its not, and skip that node.

node type (so, d710 becomes d710-vm).

include interfaces that link to the layer 1 switch.

have slices.

existing, and skip sending email.

* Any experiment can now be paniced; formerlly, only firewalled
  experiments could be paniced. There are new options in the admin
  menu section of the Show Experiment page. ProtoGeni experiments can
  also be paniced.

  At the moment, this is an admin-only option. Should it be?

* There are now two levels of panic. The first level (and the default)
  is to throw all the machines into the Admin MFS and reboot them.

  Level 2 (OH NO!) disables the control network. Not sure we need two
  levels, but hey, it was easy.

* Admins can swapout a paniced experiment. There is new code in tbswap
  to watch for a paniced experiment, and zaps the disks.

* Add menu option to clear the panic. Also admin-only. 

* Minor tweaks to Show Experiment when experiment is paniced.

instead.

Generate authority certificates for local sub-authorities (i.e. authorities
corresponding to a local project) on demand.

Map per-project URLs to the same XMLRPC server handled in the context of
the authority for the specified project.

Make the SA give out per-project credentials when it's asked for a
GetCredential in a sub-authority.

Turn on debug printing by default for now.

Previously, if "prepare" was set (i.e., at node reload time) we would wipe
out all the old partition info for a node before we processed each image.
The result is that only the last image loaded would be recorded as "on disk".

Lots of minor stuff, interfaces, command line arguments (cleaned up documentation), global rspec attributes, minor node attributes, location tags, bugfixes, etc.

Basic link lookup and interface lookup is complete. Need to finish bandwidth
calculations, link types, wireless links, and interface printing.

How simple you all seemed compared to the terrible links ahead...

This also adds the disk_image tags for rspecs.

Still remaining: widearea nodes, links, osids, single-node mode.

Basic framework is in place. Availability and printability is codified into
functions now. Node.pm modified so that lookups can be done in bulk for
efficiency. When making XML, we now create XML trees rather than printing
out XML fragments.

This only copies over existing schemas and does not build them.