Reference commit nexedi/erp5@455b24e5

 This change follow changes on ERP5 code base:
  nexedi/erp5@4b16e1dd

  By invoking validate, the interaction workflow will be kicked and allocation will be triggered. In order to control the invocation order on the test, use portal_workflow API.

  This test is failing as it is incompatible with the latest version of renderjs. The Monitor UI is going to be merged into RenderJS, so keep this test as an expected failure until this test be moved along with monitor implementation to ERP5/OfficeJS repository.

Person.getTitle do not contains patches for a while so we fall back on default the implementation of ERP5. This don't require mantains a test on project scope.

This will prevent mistaken information when display the used amount of instances.

Move manager.instance to the part of code which doesn't get preempted 

/reviewed-on nexedi/slapos.core!21

The logic implemented on client is (for started instances):

Notify instances are available to master
run promises if fail report error
Notify instances are started to master

The change on this commit prevent us to consider the notification of availability as "no error", as promise can still fail, so we just ignore and let the decision to suspend for the next run.

This makes the ease to get the message, on certain special cases.

See ERP5 commit 8995b64508f6d7c0fee1ef475de3dc3145ddc84d.

  This version include initial plugin system and cgroups integration.

  We also included a significant change for provide the partitions network informations, which now
  don't require to come from slapos master anymore, so it is possible to make local extensions.

This report list the usage of resilience instances per user, and give the % of the adoption.

This report lists the Computers that host at least one instance of a cluster which is returning error (see Related ticket columns). This list also the utilisation ratio of the partitions, and as well the percentage of the partitions inside which are reporting errors.

This report aims to help on Support Decision Process to eliminate the biggest amount of errors.

-  Move all Computer attribute initialization into Computer constructor
-  Rename 'construct' to 'format' because we are in slap.format
-  Introduce Manager base class
-  Implement CGroupManager to take care of cgroup's cpuset

-  Add TUN class
-  Check Interface is a bridge before performing bridgey operations
-  Replace outdated 'tunctl' binary with 'ip tunctl'

  This raise my lead to a significant change on the API, which
  may require to refactor several tests, to this change keep s a
  backward compatible API.