GitLab

vs. not on the control net.  The "media" command wasn't implemented in the
Linux dhclient-script anyway.

Stargate client-side boot support.

Similar to linux, but the sg linux distribution is based on debian, so some
things are a little different.

WARNING: if you do a client-install in a cross-compile environment
(with DESTDIR set), watch out for bogus symlinks in <root>/etc/rc*.d!
Have to fix these up by hand in this case.

Also note that the stargate rc script doesn't yet start the robot pilot
daemon - Tim said he needed to do finalize some stuff w.r.t. this, so
he would take care of modifying the rc script.

The ntp settings are still not ideal.  Even though the root fs is on a
journalling jffs filesys, its still not good to write to it regularly.  Need
to point the ntp drift file off into /var (ramfs)

Cross compilation fixes for the stargates, 'gmake client' should now
build, link, and install properly.  Haven't really tried to run stuff though.

	* GNUmakerules: Add target for stripping executables, used instead
	of "install -s" since that doesn't work for cross-compiling.

	* Makeconf.in: Add ELVIN_CONFIG variable that refers to
	'elvin-config'.

	* configure, configure.in: Detect and save the elvin-config path
	since we need a different one for cross-compiling.

	* event/lib/GNUmakefile.in, event/link-agent/GNUmakefile.in,
	event/linktest/GNUmakefile.in, event/program-agent/GNUmakefile.in,
	event/proxy/GNUmakefile.in, event/tbgen/GNUmakefile.in,
	event/trafgen/GNUmakefile.in, os/dijkstra/GNUmakefile.in,
	os/syncd/GNUmakefile.in, sensors/slothd/GNUmakefile.in,
	tmcd/GNUmakefile.in, tmcd/linux/GNUmakefile.in: Cross compilation
	fixes, don't statically link on arm, create "foo-debug"
	executables with debugging info and install separately stripped
	ones instead of passing "-s" to install.

	media "media 100baseTX mediaopt full-duplex";

since they will be in autosense mode by default, which does not play
well with switchports that are fixed to speed/duplex.

file /var/emulab/boot/runlilo exists.  This is used for post-diskload
configuration.  The rc.frisbee script modifies the lilo.conf file if
necessary (for disk types other than IDE) and touches the runlilo file.
So next time Linux boots, it will re-run lilo and get everything in synch.

So how do we boot Linux the first time before lilo has been rerun?
I'm glad you asked!  rc.frisbee also runs the magic groklilo program which
knows how to set the one-time command line in LILO.  We use this to set
"root=802" or whatever to make it boot that first time.

   No longer rely on looking at kernel boot time messages and extracting
   a hardware signature to determine the nodetype to then determine the
   control net.  Now we just DHCP on all interfaces and decree that the
   interface that answers is our control net interface.

   An extraordinary number of sleezy tricks were needed to get FBSD4,
   FBSD5, and RHL to DHCP on all interfaces without changing any standard
   scripts.

   For now, the nodetype/cpuspeed/chipset scripts still exist for the
   benefit of healthd, which uses the output of nodetype to determine
   what kernel module to load.  We should fix this.

   Side-effect: pump, the old RHL DHCP client, is history!  For older
   RHL releases, you will need a version of dhclient.

   Side-effect: in Linux, all non-control net interfaces are left up
   but without a legit IP address.  This is a consequence of dhclient.
   In FBSD, it was trivial to clean this up, RHL will take a little
   more work.  Up or down, it shouldn't matter.

2. Add an mfs-install make target, a scaled-down version of the client
   install.  Added a mandatory DESTDIR check so you don't accidentally
   install in the wrong place on boss.

loopback interface (yeah, Linux will really do that...)

possible to:

	gmake client
	sudo gmake client-install

on a FBSD4, FBSD5, RHL7.3, and RHL9.0 client node.

There are still some dependencies that are not explicit and which would
prevent a build/install from working on a "clean" OS.  Two that I know of are:
you must install our version of the elvin libraries and you must install boost.

1. "make clean" will just remove stuff built in the process of a regular build
2. "make distclean" will also clean out configure generated files.

This is how it was always supposed to be, there was just some bitrot.

of their file server (aero)

Extend DISABLE_NAMED_SETUP: when set, we don't set a nodes hostname to
<name>.<eid>.<pid>.<domain> since that won't resolve.  Just stick with
pc<XXX>.<domain> in those cases.  The various sethostname* scripts are
now .in so that they get preprocessed to check for the option.

the top level.  This will build all the necessary binaries and then install
them.  This works on FBSD4 and RHL7.3.  It still doesn't work on FBSD5
(newer compiler that no longer supports a style of use of _FUNCTION_ in the
event lib) or RHL9 (event lib needs SSL lib which has a bad dependency
on Kerberos).  Notes:

- requires that elvin libraries be installed on nodes (they are) to build
  event agents, requires linuxthreads be installed on FBSD (it is now) to
  build imagezip (which is installed, but is not strictly necessary)

- installed event-agents and other binaries are stripped

- added a few missing files to the source tree for bsd (healthd.conf)
  and linux (healthd.conf, rc.local)

- the only thing that doesn't get rebuilt in /usr/local/etc/emulab is
  healthd, I couldn't quickly find how it gets built

- uses a scaled down version of libtb with no DB functions (since mysql
  isn't installed on nodes).  N.B. DO NOT DO A CLIENT INSTALL FROM YOUR
  REGULAR OBJ TREE OR ELSE YOU MAY WIND UP WITH A NEUTERED VERSION OF
  libtb.a!

The build-as-well-as-install semantics are counter to the regular install
targets, but this is what we gotta do for now.  Once the TB source builds
under Linux and newer BSDs, we could undo this and just require that people
do a regular "make" followed by "make client-install"  OTOH, there should
be no reason to require installation of mysql and other server-side packages
just to build clients (or make them sit through the compilation of assign),
so maybe we will keep the client build special.

Sometime we might want to offer an option on the experiment create page to
automatically setup the 4th partition for people.

For now they can run it as /usr/testbed/bin/mkextrafs

passable manner. Mostly this change overwrites Mike's changes for
dealing with the pump race. Rather than setting ONBOOT="yes" for all
interfaces (which is what triggers the pump race), only set the control
interface to ONBOOT="yes", all the others get ONBOOT="no" so that
ifup will not be called on them. Also remove the hostname set from the
ifcfg scripts since that totally messes up dhclient-script, which is what
is used in Redhat 9.0 (instead of pump).

The problem with ONBOOT="no" is that the interfaces will not get
probed if ifup is not called. This messes up testbed ifconfig later.
So, added an rc.linux script that does just the probe for all interfaces.
This script is called from rc.bootsetup, if rc."osname" exists.

this was to add soft reconfig support so that nodes could be
reconfigured without having to reboot them. This appears to work, and
has been tested with jails getting moved around. I've also tested the
new code on the MFS, but still no testing has been done on PLAB nodes.

The main change is that most of the code moved out of libsetup.pm, and
was split into constituent rc scripts, each of which does its own
thing, including cleaning up and preparing for making an image. Most
of that central knowledge has been moved out into the scripts. Still more
to do but this was a good start.

* Download the eventkey with new tmcd call.

* Pass -k option to various agents so that they can verify the HMACs
  in the incoming notifications.

* Change program agent; The list of agents from tmcd now includes the
  command, which is written to a config file for the program-agent to
  read in. The command string in the event is now ignored.

* Build the local proxy for linux, and add the goo to start the local
  elvind and use the proxy. It has been this way on FreeBSD for a
  while, but I never got it installed for Linux before now.

  code that was in libsetup has moved into this library, and underwent a
  giant cleaning and pumping up. The interface from your typical perl
  script now looks like this:

  use libtmcc;

  if (tmcc(TMCCCMD_STATUS, "optional arguments", \@tmccresults) < 0) {
      warn("*** WARNING: Could not get status from server!\n");
      return -1;
  }
  foreach my $me (@tmccresults) {
	print "bite $me";
  }

  The arguments and results are optional values. There is a fourth optional
  value that is a hash of config options (basically converted to command
  line switches passed to tmcc). For example, to set the timeout on an
  individual call, pass a fourth argument like:

	("timeout" => 5)

  There is also a way to set global options so that all subsequent tmcc
  calls are affected:

	configtmcc("timeout", 5);

  I'll probably clean this up a bit to avoid the direct strings.

  The result list is a list of strings. Since we are trending away from
  using tmcc to transfer large amounts of data, I think this is okay.

* A new tmcc.pl which does little more than load libtmcc and use it.
  This will become the new tmcc, with the existing C version becoming a
  backend binary for it.

* All of the perl scripts in tmcd have been changed to use the new
  library. I left the few uses of tmcc in shell scripts alone since they
  were of the simple variety (mostly "state" command).

* And again, if you have read this far, you will learn why I bothered with
  all this. Well, the existing code was really bad and it was getting out
  of control. Sort of like a squid that was getting harder to control as
  its rotting tenticles slithered into more and more scripts. Anyway ...

  More important, my goal is to use the libtmcc library to add caching.  I
  have not worked out the details yet, but I am envisioning a configuration
  file, perhaps generated initially by tmcd, of all of the config
  values. If the library finds that file, it sucks the info out of the file
  instead of going to tmcd. Eventually, this config file would be generated
  as part of experiment swapping and stored in the DB, but thats a longer
  term project, and perhaps orthogonal (how we fill the cache is not as
  important as adding the ability to use a cache, right?).

  Note that certain operations (like "state" and "ready") are flagged by
  the library to always bypass the "cache".

inside, but rather I do just enough to get the card booted (using
Abhijeet's minicom/expect scripts, but with changes to support the
configuration coming from tmcd. I also create a file of interface and
routeadd directives, so that the network configures properly, but
thats about it. Getting a more complete client side environment that
includes perl and sshd for the arm will have to wait.

Here we have the Linux delaysetup startup script - first revision.  This
uses iproute2+tc, modprobe, and iptables to setup traffic shaping.
A few notes are warranted:

1) [g]red is not yet supported
   - need to make these modules classful in the kernel first

2) sysctls are used here to up the amount of buffer space available for
   sk_bufs (socket buffers).  Couldn't see a place to do this in the kernel
   config.

3) Only linkdelay support is implemented
   - probably could add normal delay-node support without too much trouble.

4) reverse pipe numbers are (currently) ignored
   - not needed since the IMQ device used to shape incoming traffic
     is distinct from the actual interface - no namespace collision.

5) Kernel selection is similar to FBSD: check running kernel, and reboot
   if the kernel version isn't what we expect (have to rerun lilo too)

* New modules.conf that works with 7.1 and 7.3 (old one broke on 7.3)
  - Uses module aliases, instead of probe directives

* Updated sfs-install target for linux startup cmds
  - RCDIR variable was used inconsistently
  - reorged rc.d startup script links to not start sfs by default
  - still need to revert back to the 6.0 startup script in the cvs repo

* Updated common directory "other" target to include slothd

move the /etc/cron files off so they are not run.

we make, into these makefiles cause we tend to forget things! Also
chain to other makefiles as appropriate, including those in the event
directory and elsehwere to install client side programs and the like.

the different node types

on the RON nodes and in the new widearea image.

as a script in init.d that gets called when the node goes into
runlevel 6 (reboot.)

(Note, it has to get 'started' earlier, when going into runlevel 3,
because if it has not created a lockfile in /var/lock/subsys,
it will not get called when shutting down.)

Removed rc.route script which is now generated automatically by libsetup
doroutingconfig code.

Random: removed some unused variables from freebsd/liblocsetup
(IFACE, CTLIFACENUM, CTLIFACE) which collectively hardwired the control net,
probably haven't been used in a while.

Client from rc.setup. Send the ISUP event when testbed configuration
is complete.

than having node types in interface-specific configuration scripts, they
use control_interface so that this information is only stored in one
place.

Also added ifcfg-eth{1,3} scripts which are not used now, but make it easier
to handle differing control interfaces in the future.

interface for this node.