GitLab

Using "set-rwclone" ala:

    set $bsobj [$ns blockstore]
    $bsobj set-lease "emulab-ops/bar"
    $bsobj set-node $node
    $bsobj set-rwclone 1
    ...

in your NS file will create a clone of the indicated persistent blockstore.

Somewhat limited in utility since you can only have one clone of a
particular blockstore per experiment.

Sets PERMS to one of RO, RW, or CLONE. Note that currently CLONE is only
seen by a blockstore VM. The node attached to the blockstore still sees RW
(to avoid backward compat problems).

release the physical node before the virtnodes on it. This has not been
a problem before cause physical nodes are not an actual sliver, it is
allocated internally by libvtop.

need to pass IsFeasible() a different variable.

Estimates the maximum possible extension for a slice.

change as I get further into it.

Fix for shifting labels on finalize step

Fix for issue #155 . The problem was in the ClearFormErrors function. Instead of hiding only the labels with '.control-error', all '.control-label's were instead being hidden

See merge request !15

A handful of code paths now use the new option to handle infrastructure
tasks where doing the right thing is more important than preserving
guarantees about future reservations (e.g., shuffling around holding
experiments, moving to hwdown, etc.).

nodes in the topology. Instead of binding with a component_id, implement
new variant of the "relation" element, called "instantiate_on" that
specifies the client_id of the parent node. This becomes the "fixed"
slot of the virt_nodes table entry for the Xen VM.

This is different than the traditional behavior of this defs- variable.
Previously it caused tmcd to not expose any NFS mounts to shared-host vnodes.
We relax that now to allow exposing such mounts to vnodes with routable IP
addresses.

The rationale for this change is simply that the original option was only
intended to prevent exporting mounts to hosts that could not reach the FS
node anyway due to their unroutable cnet IPs.

Previously, we would not pass the mounts via tmcd but they were still
exported from fs.

via the -c and -E/-F options:

* On normal nodes, we used to use evproxy to capture experiment events
  on ops and forward them to the node local pubsubd (where all local
  agents connect to). Now we restart pubsubd with the -c and a -F option
  that says to subscribe to all events '(EXPT == "pid/eid")' at the
  parent pubsubd on ops (-c event-server).

* On shared hosts, we used to start kill pubsubd and start clusterd. Now
  we restart pubsubd with just "-c event-server" which causes it to
  subscribe to all events at the parent pubsubd on ops. This is how
  clusterd did it as well.

Another change is that instead of just kill pubsubd and starting with
new options, we now write a pubsubd.conf file into /usr/local/etc. The
rc.d startup files look for this file and source it as a typical .conf
file and pass the options to pubsubd.

I added deletion of these file to the prepare script.

eventually be used for pubsubd clustering using subboss, but that is
a future commit (read: when I have time to get back to it).

service, which implements real progrom agents in the CM. See
commit 760e5d43.

  <services>
    <emulab:program-agent name="prog0"
         command="/bin/ls -lt &gt;&amp; /tmp/foo"
         onexpstart="true"/>
  </services>

Much like a normal "execute" service, these are event agents instead
of boot commands (which run on every boot of the node). When onexpstart
is true, the command is run at event time=0 only. All program agents
are set up to respond to tevc commands like program agents in an classic
Emulab experiment.

not really matter before, but with new program agent support, we have to
start it after nodes are "ISUP". This is going to take a little more
work to get right, not sure if Restart() is going to work right?

a transition. Also, call HasStartupAgent() to see if we have a program
agents with onexpstart=true; we count these as startup "execute"
services.

services section. We do not support timed start, they can only be
started at time=0 via the "onexpstart=true" flag.

completed. I used a dismiss modal. Also clear the Days box.

Just to be safe, safe, safe.

These routines map from the event constant strings to their numeric
equivalent.

Make all arguments to GetEventState() optional.  HasStartupAgent() look
in the program agent and event list, for programs that have been
scheduled to start at time=0.

the Geni path. Typically this is done in libvtop after node allocation,
but we call the mapper from getticket with alloconly set, and we end
up skipping that. I have a note in libvtop to look at this, but I want
to get a fix in before the push to stable.

Have not switched to using it yet, must update all the FreeNAS boxes first.

already allocated nodes to be assigned to the reservation.

to not finish.

extra double quotes around it. Kill them so the instance lookup does not
fail. Someday I will figure out why this is happening.

* Move user mod (gecos,password) into the accountsetup proxy instead of
  ssh chpass. Wrap all usermod/chpass system calls in a loop that looks
  for the busy file error, back off and try again for a while.

* Add same wrapping to local (boss) calls of usermod/chpass. I put that
  function into emutil.

* Rename old modgroups in the proxy to setgroups, since that it is what
  it was actually doing.

elabinelab experiments cause it messes up geni rack builds. Revisit this
later, but elabinelab is mostly a Utah thing, and we already block port
111 at the firewall.