[BIGTOP-1912] possible bug for spark-slave?

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: backlog, 0.8.0
• Component/s: vm

Environment

centos 6, openstack

• Assignee: Unassigned
• Reporter: Minying Lu
• Created: 27/Jun/15 16:30
• Updated: 27/Jun/15 16:30

Description

I saw this error when I'm running vagrant-puppet on openstack (see BIGTOP-1911 for patch detail)
Spark worked on master node, but not on slave node. It might be a puppet problem or a networking problem because spark randomly choose ports? The fact that it worked in master node made it looks like spark-master is doing some configuration for spark-slave and when in the slave node spark-master is not running, spark-slave breaks.

I set /etc/hosts with ip address and fully domain name of the host

==> hadoop-bigtop2: Error: Could not find resource 'Service[spark-master]' for relationship on 'Service[spark-worker]' on node hadoop-bigtop2.csail.mit.edu
==> hadoop-bigtop2: Error: Could not find resource 'Service[spark-master]' for relationship on 'Service[spark-worker]' on node hadoop-bigtop2.csail.mit.edu
An unknow error happened in Vagrant OpenStack provider