* wip(nodes): do not try to run iptables code on down nodes

issue #882

* test(nodes): test machine set down if node dies

issue #882

* wip(nodes): check node is down before using machine

- set the machine to down
- do not try to use the node connection if it is down.

issue #882

* test(nodes): domain in deleted node should be down

issue #882

* test(nodes): check operations after node down and up

isue #882

* wip(nodes): do not cache active node if forced

issue #882

* test(nodes): check status in dead nodes

issue #882

* wip(test): check id vm

issue #882

* wip(nodes): connect to local node if remote down

issue #882

* wip(nodes): return cached state from frontend

issue #882

* wip(nodes): shutdown all domains of a node

issue #882

* wip(nodes): make sure the node is still active

issue #882

* refactor(backend): show trace errors when failed

This is useful now we are in development. This should
be removed once it reaches production level.

issue #882

* test(nodes): we can't enforce same VM

removed this test. Slightly improved remote clones checks

issue #882

* wip(tests): make test disks small

issue #882

* test(nodes): test active clones in disabled node down

issue #882

* feature(nodes): cope with dead nodes

- Do not search machines in disabled or down nodes
- Set machines down in disabled nodes
- Refresh machines status on dead nodes

issue #882

* wip(test): reviewed already started tests

issue #882

* refactor(nodes): removed debug

issue #882

* wip(frontend): show nodes enabled/disabled buttons

issue #882

* wip(frontend): request check vms after adding new node

issue #882

* wip(nodes): fix problems when nodes suddenly down

set node status inactive if found down

issue #882

* wip(nodes): do not use down nodes

issue #882

* wip(nodes): set node inactive if ssh fails

issue #882

* wip(nodes): search domain in node only if active

issue #882

* wip(nodes): previous VM can be used on opening domain

issue #882

* refactor(nodes); speed up balancing and listing

issue #882

* wip(backend): removed warning , probably we should mark req done

issue #882

* wip(tests): removed unnecessary checks

issue #882

* wip(nodes): fixed wrong cached inactive call

issue #882

* test(nodes): force node to start in

issue #882

* wip(nodes): non-clones go to main node

This way when it is prepared as a base all the disk
volumes are here. It would be more difficult to handle
creation of bases that are in a remote node.

issue #882