not just the experiment creator.

imaging (cause of a failure in the backend).

there for 30 minutes or so.

  -- NOTE: These are mostly pulled from ipmitool.  Other protocols may need
  -- other values or not care at all, hence the addition of OTHER to the 
  -- list and a default of NULL.

Uses the tmcc/ubuntu15 os-dependant directory, which seems to work.

For the case where zapdisk doesn't exist.

also support alterntive ipmi privilege levels

Also need to add support for separate ipmi encryption keys from user
passwords, but it turns out that the current db schema does actually
support that already and at least one script more or less does it
already (power_ipmi.pm), though another doesn't (power_ilo.pm), so we
can fix that separately there.

This is to allow support of additional read-only users for gathering
power data that don't have to also be given the admin password in form
form of the hex encoded Kg key.

Also add a field so that we don't need to assume that the privilege
level is necessarily ADMINISTRATOR.

This was causing eventual unmount of the customized FS to fail and
apparently causing the changes to get lost and/or cause FS problems.

terminates properly (zero status) instead of being killed by
the reboot/halt (and exiting with -1). Noticed on Debian 8.0.

For those of us with 80-column vision...

Awesome.

It took me so long to find this stupid bug, I don't want future me
to have to hunt for it again...

Since this is the obvious fix to libabac not working after the upgrade
to 0.1.9.

notifications to an image's origin cluster, another for the origin cluster
to record images that need to be imported from a remote cluster after
notification.

instead of doing nothing.

Did this quite a while back, but haven't tested yet.

Changes to make a client even more passive when there are blocks in
flight (requested by someone else) that they can use. The goal here
is to keep a late joining client from making requests for chunks that
all the others have already seen. This is not a big problem in the
default case where clients randomize the order of chunks they request,
but when they are making sequential requests it can be a problem.