GitLab

This was in the CENTOS63-64-STD image at APT, but not at the MS, and not
in the CENTOS66-64-STD image.  I must have added it by hand there. Now it
is installed as part of the client-side install for CentOS 6 flavors.

but only surfaced as a problem on Clemson. Odd.

not always be immediately available. Mike says there is a race in mountd.
Surprise!

Handles the case where a wire in the wires table has an 'Unused' role
(which is how logical wires begin life -- then once their underlying
physical links are built into vlans, snmpit changes their wires.role
changes to 'Node').  This change ensures undef ports (which is what is
returned for wires entries with 'Unused' roles) don't make it into the
getVlanPorts list, where they cause snmpit to die, of course.

The stuff in db/Port.pm (Port::LookupByIface) wasn't handling the MLE
case where there are two wires table entries for physical wires with a
logical wire implemented by them; for my experiment, when snmpit was
trying to create the physical vlans on the apcon, it would grab the
logical wire instead of the physical one as it populated the Port
objects, and that screwed vlan creation up (the member port for the
physical layer 1 vlan gets labeled with a switch WIRE_END instead of
node -- which is correct for the logical wire, but not the physical).
So, I changed the query to check that wires.logical ==
interfaces.logical ; see the code.  This is a limited fix, maybe --
maybe that method should really do more verification to make sure it's
grabbed the correct wires table entry -- but since we only have 2 layers
it's plenty good for now.

change had some problems.

Also some fixes to the signature code.

No one has used this code in 10 years and not one will ever use it again,
but it broke my testing...and PISSED ME OFF!

Time for a LILO revival!

I'm not sure exactly why this was originally like this (let the node out
of reloading before doing the console-based reconfiguration --- cleaning
out the old config and writing a default one, via the console), but it
now fails because the node gets released from reloading (and thus its
reservation is cleared) before the reconfiguration step can grab the
console line via tiptunnel --- and a subsequent tiptunnel connect fails,
presumably because the reload_daemon no longer has permission to grab
the line.

Anyway, we'll try this... I can't see any obvious negative side
effects.

the apt_instance table record, and display that in on the admin extension
page so that we have it all in one place instead of email messages.
Also show the cluster usage graph in on the admin extend modal.

extensions, to show on the admin extend page.

For imagezip. Since I am the founding--and only--member of both camps,
I get to decide what is right! I am back to generating imagezip relocations
(for FreeBSD images only) since those images are just a big pile of bit-slag
if you want to load them anywhere other than their standard location.

Just luuuv string manipulation in C...

instead of cloudlab-ops, which we want to use for the combined cluster
email list. Send more email to logs instead of ops, now that we are
generating enough of it.

their experiments, we fall back to naming it for them.

picking pieces out, so I added this:

  my $foo = GeniHRN->new("urn:publicid:IDN+emulab.net:testbed+slice+myexp3");

and provides:

  $foo->domain();
  $foo->type();
  $foo->id();
  $foo->urn();

and has a stringify method that returns the urn, so that existing code all
works.

The problem though is the XML::RPC and Frontier libraries provide no hooks
to catch this, and rather then using a stringification, they both convert
all blessed references into structs, and so anyplace that puts a urn into
something to go out on the wire, has to be changed to force to the string.

Damn, how disappointing! So all the code is here but basically disabled
until I find time to go through all the code.

The "Emulab configuration" of the mserver will now use uploader_path
if set rather than path.

A slight semantic change goes along with this. If the uploader_path is
set, we assume that it is a temporary file to upload to and we (the mserver)
do not do the dance of uploading to our own temporary and moving it into place.

of image files present.

Add uploader_path to communicate the path to use to write the file.
Set in create_image.

   RELOADSETUP -> RELOADFAILED
   RELOADING -> RELOADFAILED
   RELOADFAILED -> SHUTDOWN

defaults to NULL since we did not require names prior to this, thus there
is no unique key defined.