Bug #396

/tmp/one not populated if headnode can't ssh to host node

Added by Dan Yocum over 10 years ago. Updated over 10 years ago.

Status:ClosedStart date:11/08/2010
Priority:NormalDue date:
Assignee:Daniel Molina% Done:

100%

Category:-
Target version:Release 2.0.1
Resolution:fixed Pull request:
Affected Versions:

Description

In the instance when a misconfigured host/worker node is added via 'onehost add blah', an attempt to redistribute the contents of /tmp/one/ doesn't occur. Here's the scenario:

1 - host node is deployed, but authorized_keys2 isn't populated, or the permissions change, or...
2 - 'onehost create <host> im_ vm_ tm' is executed on head node
3 - the problem on the host node is fixed
4 - when the head node successfully is able to ssh to the host node, the software should be redeployed to /tmp/one on the host node.


Related issues

Related to Bug #412: oneadmin does not have permissions to touch /usr/lib/one/... Closed 11/15/2010

History

#1 Updated by Ruben S. Montero over 10 years ago

  • Assignee set to Daniel Molina
  • Target version set to Release 2.0.1

#2 Updated by Daniel Molina over 10 years ago

Hi Dan,

You have to use the onehost sync command that will update/copy the remotes in the next monitoring cycle.
You may want to decrease the host monitoring time that is set by default to 600 in the oned.conf file.

HOST_MONITORING_INTERVAL = 600

#3 Updated by Daniel Molina over 10 years ago

  • Status changed from New to Closed
  • % Done changed from 0 to 100
  • Resolution set to fixed

We have included a new feature to OpenNebula regarding this issue, you can check it at http://dev.opennebula.org/issues/412.

OpenNebula will check the remote files existence and will copy them if needed. The onehost sync command forces OpenNebula to update the remotes even though they are present in the worker node.

Also available in: Atom PDF