GitLab

It is invoked by suexec and cannot be a symlink.

Add the additional interfaces for boss.

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.

anything that looks like a password when writing the file on ops.

in the conf file.

If a project image has been "exported" with grantimage, os_load may not
be able to stat it due to unix permissions. So have os_load make a query
to the local frisbee master server to get attributes for the image. The
master server already knows how to deal with these exported images.
This query also works for an inner boss determining the size of an image
it has not yet downloaded and thus subsumes that case.

So we can configure clientside-only so that imagezip has NTFS support.

Also, nit: no need for WITH_EXT4FS since that is part of the default EXT
support.

Was using TBDB_TRUSTSTRING_NONE which is not even defined.

Also fixup clean target.

Whatever the issue was, I seem to have avoided it through a combo of updating
all the 9.0 stuff and reverting to an older mrouted port.

A change in the BSD ports causes the libpcap port to be installed along
with the "built in" version of libpcap. This caused the NS build to drag
in the wrong version of bpf.h.

These are setuid perl scripts that are run by "suexec" from the webserver.
suexec disallows symlinks so we must make these hard links to runsuid.

Also, get rid of message about "you must run post-install".

Don't run mrouted, it seems to crash the kernel.

Load ipdivert.ko at boot time, else the ipfw rule for NAT fails.

Bogus param list for AddPerExperimentSwitchStack; can't have scalar args
after a list arg.

Missing mailing list TBERRORSEMAIL in defs-elabinelab. Caused redundant
"testbed-ops" addresses in /etc/aliases.

Weed out said redundant mailing list names when creating aliases during
ops-install.

named.reload script has gone away. Use rndc instead.

The perl warnings can screw up some of our scripts which expect certain
output; e.g., users of "setsitevar -c" expect to see the value of a sitevar
and not some nagging perl message.

Don't want to remove SOUT altogether since I don't know how that will
affect the regular libtblog (which includes libtblog_simple). So I just
threw in an "if (0)" which prints to SOUT...

Perl 5.12.4 needs the same self-loader patch as 5.10.1.

Doh! Forgot to turn off the debugging in runsuid.c. Hence it wasn't
actually exec'ing anything!

Upgrade to perl 5.12 means no more "suidperl" (setuid perl scripts).
So we now have yet another little wrapper (security/runsuid.c) which
runs suid and whose sole function is to exec the perl script of the
same name in the /usr/testbed/suidbin directory. So a formerly setuid
perl script install now goes like:
  create /usr/testbed/sbin/mkproj as a symlink to /usr/testbed/libexec/runsuid
  install real mkproj perl script in /usr/testbed/suidbin/mkproj
When the setuid-wrapper is invoked under the name "mkproj" it execs
/usr/testbed/suidbin/mkproj. We could almost use sudo for this purpose
instead (see security/sudoers.in) but sudo loses one of the groups in
the group list.

/usr/include/utmp.h is gone in FreeBSD 9.0. In most places we nevered
needed it, but in the one case that did (tg source), it just used a
couple of the constants exposed (UT_*) and not the struct, so I just
hardwired values for the constants.

The usual tweakage to the install stuff to reflect yet another set of
packages!

will become a wiki page ...

wildcard the lnk speed, except when the link includes a virtual
node; in this case use 100Mb to avoid consuming all BW on GB links.

Integrated into install process. Now it is a part of the exclusive sbin club.