[9.2] [StackLight] failed to deploy node with grafana plugin

[Yury Tregubov]

Since 9.2 snapshot #533 deploy of node with Stacklicht plugins is failed:

2016-11-29 12:38:39 ERR /usr/lib/ruby/vendor_ruby/puppet/provider.rb:115:in `execute'
2016-11-29 12:38:39 ERR /usr/lib/ruby/vendor_ruby/puppet/util/execution.rb:219:in `execute'
2016-11-29 12:38:39 ERR ...fail!
2016-11-29 12:38:39 ERR Could not start Service[grafana-server]: Execution of '/usr/sbin/service grafana-server start' returned 1: * Starting Grafana Server

The plugin version seems to be influxdb_grafana-0.10-0.10.2-1.noarch.rpm

The test doesn't check the grafana plugin it self. It just installs the node with plugins to check interoperability with the Fuel LDAP one.
Test scenario is the following:
        """Check the LDAP plugin together with StackLight plugins

        Steps to reproduce:
        1. Checks users and groups for LDAP domains
        2. Download and install StackLight plugins
        3. Reset and re-deploy env
        4. Execute diagnostic command
        5. Checks users and groups for LDAP domains

        Duration ~ 90 minutes
        """

The fail occurs on step 3.

Test logs are available here: http://172.18.173.8:8080/jenkins/view/MOS-INT%209.X/job/9.x_Fuel_LDAP_no_proxy/lastCompletedBuild/testReport/mos_tests.keystone/test_keystone_ldap/test_plugin_interoperability_stacklight__1682427__/

Diagnostic snapshot is https://docs.google.com/a/mirantis.com/uc?id=0B1Crk-sAvGanVEtTa2xYaHViQUE&export=download

[Swann Croiset (swann-w)]

according to the grafana logs:
2016/11/29 12:38:27 [log.go:84 Fatal()] [E] fail to initialize orm engine: Sqlstore::Migration failed err: Error 1045: Access denied for user 'grafana'@'10.109.7.9' (using password: YES)

Something goes wrong with the database configuration

[Swann Croiset (swann-w)]

The manifest grafana_mysql.pp is not applie (it should be applied just before grafana.pp [0] ) :

grep fuel-plugin-influxdb-grafana: /var/log/puppet.log
2016-12-01 10:13:29 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: hiera.pp
2016-12-01 10:18:14 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: firewall.pp
2016-12-01 10:18:17 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: cluster.pp
2016-12-01 10:19:53 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: cluster-haproxy.pp
2016-12-01 10:20:04 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: haproxy.pp
2016-12-01 10:20:13 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: influxdb.pp
2016-12-01 10:39:09 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: grafana.pp
2016-12-01 10:39:28 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: grafana.pp
2016-12-01 10:39:43 +0000 Scope(Class[main]) (notice): fuel-plugin-influxdb-grafana: grafana.pp

It looks lika that all tasks asssigned to "primary" roles are not executed !

[0] https://github.com/openstack/fuel-plugin-influxdb-grafana/blob/stable/0.10/deployment_tasks.yaml#L203

[Stanislaw Bogatkin (sbogatkin)]

I tried to reproduce this bug at local environment but in my case redeployment went pretty fine.
I have been used snapshot 9.2 #627 for it.
Closed as invalid.