GitLab

process that hands off).

email, so that it has the correct subject tag.

case we have to ask the clearinghouse.

so that the SCS will accept it. Oh, and I learned that the SCS is
sensitive to the order of elements in the XML file! The link *must* be
after the nodes, or the SCS refuses to do anything with it. Sheesh.

when we have a candidate powder aggregate.

as per issue #440.

locally on boss.

We can probably not start ops mysqld anywhere.

pass along the IP of the FreeBSD OPS jail to XEN client setup, and havit
add that IP to the iptables rules (antispoofing) so that ops can talk to
the control network.

if we need to assign a routable IP to boss, we also need a routable IP
for the OPS VM, and the easiest way to do that is with an address
pool (count=1). Also changes to pass that IP address to the XEN
clientside code so that it alters the antispoofing firewall rules
that every XEN VM gets.

reservations and datasets; if the aggregate is marked as adminonly,
then user must be red-dot.

These scripts should not ever exit, lest they block other hook
scripts from running.  return isn't really correct
either, but at least it won't prevent future scripts from running;
it will just nullify some of the logging dhclient-script does.

Also fixes an unrelated minor bug.

We cannot simply source into the shell; must stringify values with
whitespace.

- If XEN_SLICE2IMAGE attribute is specified, create a 16GB partition 2
- Reduce the fraction of space used by the thin pool, especially on
  non-shared hosts.
- If xen-vg volume group already exists, use it. But also check to
  make sure thin-pool exists and create it if not.

The last change allows the user to specify a blockstore on the physical
host to use rather than having libvnode_xen build up its own.

deferred if its offline.

activity for the project to make sure it is exported to boss.

Avoid SIGPIPE caused by gzip exiting after finding unexpected stuff
at the end of the compressed kernel.