All nodes reported as Down w/ System Information unavailable in WebUI

[defkev]

Only CPU/Memory utilization gets plotted.

Tested on CentOS 7.1 and Ubuntu LTS 14.04.3 Both show the exact same behavior

Running utilities/contrail-status reports everything as ACTIVE

agent                : ACTIVE
agent_1              : ACTIVE
analytics-api        : ACTIVE
apiSrv               : ACTIVE
cass                 : ACTIVE
collector            : ACTIVE
control              : ACTIVE
disco                : ACTIVE
dns                  : ACTIVE
ifmap                : ACTIVE
named                : ACTIVE
query-engine         : ACTIVE
redis-w              : ACTIVE
redis                : ACTIVE
schema               : ACTIVE
svc-mon              : ACTIVE
ui-jobs              : ACTIVE
ui-webs              : ACTIVE
zk                   : ACTIVE

And will it in fact looks like all is running fine, no node information gets collected.

Tabbing through all the spawned screens shows nothing out of order.

Doing a curl -v http://xx.xx.xx.xx:8081/analytics/uves/control-node/35f2318c-6bd3-4a3c-b2e8-555bdcc7a245?cfilt=NodeStatus (where 35f2318c-6bd3-4a3c-b2e8-555bdcc7a245 is the UUID of one of my Control Nodes) i only get an empty JSON result set like:

{}

Which would imho explain the System Information unavailable in the UI

Furthermore contrail-collector.log is constantly logging a:

... [Thread 139688626226944, Pid 27565]: ParseNodeParsing Empty node

over and over again.

From what i could dig up so far i am missing a contrail-nodemgr instance for my different node types. Will the python script is in:

• /usr/bin/contrail-nodemgr for CentOS
• /usr/local/bin/contrail-nodemgr for Ubuntu

Which expects to be run as a supervisor event listener and the the supervisor config files are in /etc/contrail (supervisord_control.conf, supervisord_vrouter.conf, ...) contrail.sh doesn't really seem to use supervisor at all??

If i create a *.ini for e.g. the controller as contrail-control-nodemgr.ini in /etc/contrail/supervisord_control_files/ and start the supervisor using /usr/bin/python /usr/bin/supervisord -c /etc/contrail/supervisor_control.conf supervisor starts reporting (for example) the disk util of the node in /var/log/contrail/contrail-control-nodemgr-stdout.log but the UI still reports System Information unavailable

Anybody?