GitLab

three different kinds of datasets was a miserable failure. I backed
out part of that change.

not try to info the parent OS.

I know, I know. You are asking "How could you have possibly forgotten
about the m400 nodes!?"

until I fix up the UI.

and spit out lat/long coords of the path. Needs more work, but I
wanted to get it into perl for next time we need it.

Also whether it uses the hacky quick sort implementation.

admin when extending a slice.

services file so that we have names for the lower ports.

/usr/testbed/etc/nmap, and then use a whacky set of options
to get nmap to scan lower 1024 ports plus our port set.

Many of these have known exploits so we could use this list in a
scan to find experiment nodes that might be vulnerable.

Unfortunately this list is not quite in /etc/services format, I
ran out of gas before adding short pithy service names.
So the little app-scan.pl script is an example of parsing out the
port numbers and nmap scanning them. If this list was in the
correct format, you could use it with nmap directly (--services-db).

We currently load/reboot these like PCs, not expect-managed
allocatable switches, so this adds a custom libosload to match
existing expectations in e.g. GeniAggregate for user-allocatable
switches.

Add another random sshtb command to do a "df" on the target host
if our imagezip attempt fails.

range chart when the whole panel moves down in the stack.

the change to MaxExtension(), but need to do it in Extend() too.

make sure to clear it from the per-aggregate rows in the DB.

aggregates and skip them (like computing max extension).

back to the experiment if they come back on the route.

parent OSID, use what is actually running on the shared pnode cause it
might be different.