Documentation to follow.

Add "-Q <interval>" option to the master server to allow it to act as an
IGMP V2 querier in environment where there is otherwise not one. It does
essentially what the perl-based querier (code.google.com/p/perl-igmp-querier/)
does, sending out a v2 membership query at the specified interval.

This eliminates the need to run mrouted in some environments (e.g., elabinelab)
just to issue IGMP queries. As a result, all the boss-install and elabinelab
setup related to using mrouted to perform this function has been removed.
The elabinelab CONFIG_MROUTED option has been changed to CONFIG_QUERIER
(the former is still recognized and mapped to the latter). The undocumented
defs-* variable NEEDMROUTED has been changed to NEEDMCQUERIER (the former
still exists in install/installvars.pm.in but is always set to 0) to more
accurately reflect the variable's purpose. If NEEDMCQUERIER is set, then
the mfrisbeed startup script is modified to add the "-Q 30" option.

The implementation of the client and server "-K <interval>" keep-alive option
has been changed to directly send IGMP v2 membership reports containing the
associated MC address.

Note that the -K options have always been a hack to work-around assorted
IGMP-related misconfigurations and incompatibilities, and really should
only be used as a last resort. As implemented, they could cause the host
machine to be pruned out of other MC groups at the nearest switch since
they only report membership in the frisbee MC group. With the master server
acting as an IGMP querier, instances of the frisbee server on that host
should no longer need to do keep alives. We still have one case where it
is needed on the client-side: a FreeBSD 8.x or later host connected to an
IGMPv2-only switch. It appears that the IGMPv3 implementation added in
FreeBSD 8.x always sends v3 reports, even when the default is configured
(via sysctl or even recompiling the kernel) as v2.

This is necessary for a "make install" to succeed without a previous
"make [all]".

The basic operational model is as follows.

* We turn the nodes on.

* Since there is nothing on the disks, they will fall through to
  booting from the PXE and will boot the newnode MFS. They all check
  in.

* We run Jon's script that adds the nodes. They are now in hwdown,
  still nothing on the disks.

* We run my script, which is driven from a datafile we are supposed to
  get from HP. This script has the ilomac, ilopswd, control mac. I
  will add another column initially; the permanent IP to assign to the
  ilo. This script does:

 + Reads the datafile to get all the stuff.
 + Reads the dhcpd.leases file to find the temporary IPs of the ilos.
 + Finds the corresponding nodes in the DB.
 + Sends over an XML file that does the following:
     - Add the elabman user.
     - Add local root's dsa pub key to the new elabman user.
     - Add Utah's root dsa key to the Administrator user
     - Sets the power on mode to auto (so that the node turns on!).
     - Sets the idle timeout to 2 hours.
 + Sets the bootorder so that PXE is first. This has to be done
   with ssh and some expect stuff I culled from power_ilo. Sigh.
 + Calls out to another script that adds the ilo interface to the
   DB (this is the management_iface script I did last month).
 + Sends another XML file that tells the ilo to reset itself, so that
   it picks up its permanent IP address.

* Now we can free the nodes from hwdown.

This isn't just Mike being anal, the presence of the tg2.0 distro actually
breaks the build.

Also take tg2.0 out of .gitignore.

localize_mfs, the later so we can dhcpd to listen on other interfaces.

the os subdir into the clientside/os dir.

Mike, use the --follow argument to git log to traverse across the mv.

Add LINUX_FSNODE to indicate that the FSNODE is running LINUX.
This is just a partial implementation, intended to support some
changes contributed by Maarten Dewispelaere at intec.ugent.be
Note that these changes do not work with windows enabled, so I added a
check for that.

This is not as general or as useful as you might think.
Right now it simply controls what device gets used as console when
customizing the MFSes. "sio" is the default. Other choices correspond
to pxeboot variants: sio2, vga, null. Choosing vga also sets the magic
VGAONLY setting in the frisbee MFS ensuring that any customized (via slicefix)
FreeBSD image doesn't try to use the serial port as console.

A real solution for choosing node consoles would not be per-testbed.
It would be per-node-type and per-node. Right now the console type is
selected in that fashion via a combination of the pxe_boot_path and
frisbee/admin MFS OSIDs. At some point we should make the console setting
explicit.

in which it will later run. Now setting up and protogeni during the
install.

* Split up boss/ops/fs install into indvidual modules; generally, what
  was a toplevel phase in the original files is not a file. This
  allowed for better code/variable reuse. No longer monolithic, which
  makes it easy to test and rerun parts.

* Incorporate "update" into the install process. Certain phase file
  can be used in update mode, as when the IP/subnet/domain changes.

* Moved the MFS setup from rc.mkelab into the normal install process.
  Users no longer have to do this themselves. Good thing.

* installvars.pm is a new library that has the merged set of the
  zillion variables that were at the top of boss/fs/ops install.

different versions into v1/v2 subdirs, since I did not like the
naming.

Add new version of the ops httpd.conf file, again taken from Dave's
linux-port branch, and then whacked on.

Add utah versions of the boss/ops config files and a utah target.
These will probably get removed later when we switch over, but I do
not want to lose the work I did to merge in the local hacks.

which is a submodule. Bail if out of sync.

This will most certainly annoy Jon, but everyone else should always be
running in sync.

view of what Portal support does:

https://users.emulab.net/trac/emulab/wiki/Portal

Notes:

* New DB tables to store the list of peer emulabs, and exports info
  for users and projects.

* Backend script to manage user and projects exports.

* Cross site login from the portal; users do not login into the peer
  sites directly.

* New portal_daemon to run on the portal, to handle exports and
  updates.

That directory has been merged with imagezip/extfs.

variables for easier config in apache and protgeni code. The port
defaults to 12369, RPCNAME to $WWWHOST, and the URL is derived from
those two.

* The httpd.conf file I started with came from Dave's linux-port branch,
  and subsequently whacked for FreeBSD installation. Sorry for not using
  git to bring the base version in.

* Configure changes to determine what version of apache is installed, and
  modify behaviour in makefiles accordingly.

* Along with Apache 2.2 comes the latest version of PHP5, and that requires
  a bogus timezone directive in php.ini to prevent endless warnings. So I
  moved the entire php,ini install from ops/boss-install to here.

Note that I had to use the 8.2 ports tree to build this stuff, and it the
usual headache cause options and directives have changed.

rspec-emulab directory (needs reconfigure) and missing rspec-geni
directory (needs git submodule init/update).

tracking with a git submodule; Move all of the makefiles into a shadow
tree (rspec-emulab), and use VPATH magic (via MOSTLY_SRCDIRS) to
install from the rspec subdir.

1. Support for "one-shot" PXE booting ala the one-shot osid. Switches to
   pxelinux to boot WinPE and then switch back after done. Painful now
   because we have to HUP dhcpd everytime we change the PXE path, but we
   may be able to fix this in the future by going all-pxelinux-all-the-time.

2. Added pxe_select, analogous to os_select, for changing the pxe_boot_path
   including the one time path.

3. Added the WIMRELOAD state machine to shepherd a node through the process.
   Still has some rough edges and may need refining.

    
    OPSVM_ENABLE=0
    OPSVM_MOUNTPOINT=""

clientside requres it to be there. As noted in previous revision, the
full emulab build always includes the clientside subdir, but a
clientside build will not necessarily include the rest of Emulab.

directory.

library.

(cherry picked from commit 940c2a0c)