GitLab

of the aggregate_history table.

We were not processing the timeout queue because we got stuck forever in
the loop that processed events. Now before looping back to sysread, make
sure there is something to read so we don't block.

When we startup or re-read the DB state, ignore really old state timeout
values; e.g., for nodes that have been dead for ages but happen to be in
a state such as SHUTDOWN that has a timeout.

In the main loop, handle any re-read of the DB state before testing the
queue length to see if we can do a blocking poll. Re-reading the state may
add timeouts to the queue.

spit to the new file. Turned out to be trickier than I thought it
would be.

useful when a tag reservation fails.

properly. Now able to allocate and redeem tickets at PLC via cooked
mode.

lanid or a vname.

encapsulated link; the vinterfaces were not getting updated properly.

I am not happy with the change mentioned above, but I needed it to
make vlan tag pre-reservations work; the encapsulation vlan has a
generated name, and that is hard to match up with the tag that is
reserved before the experiment is created, as from the protogeni
interface when stitching. I plan to come back and redo this after the
GEC.

(something I just managed to remember) is that if you call the shell
without STDERR=2, then the shell has no place to send STDERR. The
only way to get this correct is to use POSIX::dup2 instead of the
standard perl open function. Well, I bet there is a way (or 10 ways)
to do this, but this is one that stumbled onto. libaudit does a lot
of descriptor messing around.

BTW; this is also why we sometimes lose STDERR from the XMLRPC server.

This reverts commit c9d86132.

Til we resolve local multicast issues.  This should be a sitevar, but
don't have time to do all the plumbing, update scripts, etc. right now.
Also not sure if the libosload_new code is right, should the variable
I introduced by global like MAXRETRIES is?

... to make it easier to see which nodes are available for
pre-reserving to a project.

Incedentally, 'avail' has the worst argument processing ever.

GeniCMV2::CreateSliver now forks a child to to the startsliver.