GitLab

we had 768K of real memory left, but we wanted 1024 for a VM.

both sides of the wires table for external network nodes.

  <rs:site_info
      xmlns:rs="http://www.protogeni.net/resources/rspec/ext/site-info/1">
    <rs:location
        country="US" latitude="40.768652" longitude="-111.84581"/>
  </rs:site_info>

Eventually replace per-node info if its just duplicate, but for now leave
the old location in place.

Previously ptopgen was ignoring permissions granted to individual users
for OS access.  This commit also updates subos enumeration (previously
ptopgen was not looking at the permissions table at all for suboses).

   <rspec_routable_addresses
      xmlns="http://www.protogeni.net/resources/rspec/ext/emulab/1"
      configured='XXX' available='YYY' />

will correct error (no_bandwidth instead of no_connect).

Quite simple actually. In ptopgen we add all persistent stores, which
are identified by having a non-zero lease_idx. To ensure they are only
matched up with the correct virt_blockstore, we add a feature "bs-lease-NN"
where NN is the unique lease index with weight 1.0.

In vtopgen, when a request is made to map a persistent store, we first
verify that the experiment pid matches that of the blockstore's lease
(this is a fer-now simple check which is actually redundant as we make
this check in the parser too) and ensure that the blockstore is not already
mapped (indicated by the blockstore_state entry having size==0) and then
we emit an entry with desire "bs-lease-NN" and weight 1.0.

Most of the file changes were actually to get the "preassign" flag to
vtopgen forwarded into the various libvtop modules so that I can check it
and not perform the permission/size check if it is set.

Also, multiple ports on the same contact node will now be consolidated into one stitch:node tag with multiple stitch:port children.

On shared nodes, look for the 'xen-host' or 'openvz-host' osfeatures when
determining sliver types.

dom0mem setting for the node, since that varies and reduces the
amount of guest memory available.

Utah has rewired some of the its pc3000s to a procurve, which supports
vlans > 1024. Since we do not have link features and desires, we need
to trick assign by using a different link type on the wire, in both
the ptop and vtop. The vtop looks for a vlan reserved tag greater then
1024, and changes the linktype to ethernet-highvlan. In ptopgen we
look for an attribute on the switch that says it can do highvlans, and
add that linktype for any wire attached it.

Doing this required adding columns to the virt and physical blockstores
tables to mark the attributes that will be considered for mapping.
Unmarked entries just flow through to the client-side.

This commit also introduces filesystem support in the form of passing
through a mount point to the client-side.  It is left to the client to
decide what filesystem and fs options to use to setup the space, including
any logical volume aggregation required to support the request.

amount a physical has is stored in the node types table, and the
per-vm memory requirement is stored in the nodes table. ptopgen
adds up usage, and subtracts from the total for the ptop file.
The vtop number comes from a virt_node_attribute table, and we
pass this through to the client side. Note that this is less
important for openvz, more so for XEN.

In the NS file:

	tb-set-node-memory-size $node 1024

Number is in MBs. The mapper defaults this to 128 for openvz and 256
for xen. Maximum is hardwired to 256 and 512 respectively. Need to
think about a good way to configure this in.

In the node_attributes table, add an "allowed_projects" entry with
the value a comma separated list of project names that may use
this node.

I hope we don't have another way to do this; could not remember if we
did already!