Yes, out of the blue and off the wall. But I got tired of trying to
guess what we had Linux and FreeBSD use. I was surprised to discover
that we were using UDP on Linux (which caused Clemson CloudLab to fail
because they have jumbo frames enabled on their control net switches
but ops had the MTU set to 1500).

Anyway, here it is. The default setting is UDP for backward compat.
We should probably set it to TCP nowadays. There is also an 'osdefault'
setting which says use the default setting on the client OS.

at the next extension, to avoid user aggravation. Also show the reason on
the termination modal if the experiment has been locked down.

parameterized profile. Note that we don't use this path much, but
we might in the future.

account, do not create a new shadow account.

experiments. Lockdown is implemented at the portal and at the backend
cluster (ignore by cm and sa daemons). At present, the only policy is
that admin extensions longer then 14 days, set the lockdown.

job of limiting overuse of shared nodes.

error in creating it.

It's clear from the various Force10 switches and firmware versions
that we've seen now that the vlan membership PortSet bit
vector size can vary widely.  This variation even happens between
different versions of firmware on the same switch!  So, now we
detect this dynamically when the backend force10 snmpit object is
instantiated (in readifIndex()).

Required updates to Makefile-linux.sa makefiles which are used
in Ryan's build environment.

config variable that defaults to 12, but is overridden to 24
on Apt, Cloud, and the MS.

are granted for free. Better then hardwired to seven in the code, and the
new code treats zero as, no free extensions for mere users.

even if it already was (which causes it to lose the current set of vlans).

Changed mountd to write the current timestamp into /var/run/mountd.ts
whenever it has finished processing all exports files. So someone who
HUPs/USR1s mountd can check and see when the timestamp changes.

We use this in exports_setup.proxy to ensure we do not return before
mountd has completed parsing and effecting changes. Since I am too lazy
to check the before and after values of the timestamp, I just remove that
file just before signalling, and then wait for it to reappear to signify
that mountd is done.

Hopefully will prevent catastrophic DB failure as we no longer make
tens of thousands of DB queries to find the first available uid.