4

I setup Cassandra, OpsCenter and the needed DataStax agent on my EC2 Amazon machine. At the moment it's only one machine.

Everything seems to be running fine, except the node list is empty and so are the keyspaces in the Opscenter. The cassandra, datastax and opscenter logs show no errors and I followed the installation / configuration carefully. Then tried all the suggested fixes.

My guess is the problem lies in the communication between the agent and opscenter.

enter image description here

After a while these requests fail:

enter image description here

etc/cassandra/cassandra.yaml: (simplified)

cluster_name: 'CassandraCluster'
    seed_provider:
        - class_name: org.apache.cassandra.locator.SimpleSeedProvider
          parameters:
              - seeds: "1.2.3.4"
listen_address: 1.2.3.4
rpc_address: 0.0.0.0
endpoint_snitch: Ec2Snitch

etc/opscenter/opscenterd.conf: (simplified)

[webserver]
port = 81
interface = 0.0.0.0

[authentication]
enabled = False

[stat_reporter]

[agents]
use_ssl = false

var/lib/datastax-agent/conf/address.yaml: (simplified)

stomp_interface: 1.2.3.4
local_interface:  1.2.3.4
use_ssl: 0

nodetool status output:

Note: Ownership information does not include topology; for complete information, specify a keyspace
Datacenter: eu-west_1_cassandra
===============================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
--  Address       Load       Tokens  Owns   Host ID                               Rack
UN  1.2.3.4       2.06 MB    256     100.0%  8a121c12-7cbf-4a2a-b111-4ad111c111d8  1a

Nothing really strange shows up in the log except for the repetitive occurence of the following line in the agent.log:

 INFO [install-location-finder] 2015-03-11 15:26:04,690 New JMX connection (127.0.0.1:7199)
 INFO [install-location-finder] 2015-03-11 15:27:04,698 New JMX connection (127.0.0.1:7199)
 INFO [install-location-finder] 2015-03-11 15:28:04,709 New JMX connection (127.0.0.1:7199)
 INFO [install-location-finder] 2015-03-11 15:29:04,716 New JMX connection (127.0.0.1:7199)
 INFO [install-location-finder] 2015-03-11 15:30:04,724 New JMX connection (127.0.0.1:7199)
 INFO [install-location-finder] 2015-03-11 15:31:04,731 New JMX connection (127.0.0.1:7199)

To supply all the info here are the logs:

  1. opscenterd.log
  2. agent.log
  3. cassandra/system.log
Gerard
  • 3,108
  • 1
  • 19
  • 21
  • Is OpsCenter running behind a proxy at all? – Peter Halliday Mar 11 '15 at 19:49
  • No, not behind a proxy. The web interface is listening on the public ip, and the somp interface is the private ip. Is this incorrect? – Gerard Mar 12 '15 at 19:57
  • Those /tcp failures almost always point to a webserver configuration that relates to a reverse proxy. Can you provide the url in which you access OpsCenter (ip omitted)? I think the stomp interface can be a private IP, as long as the OpsCenter machine can reach it. – tom.alexander Mar 12 '15 at 21:20
  • 1
    Turns out it was AVG antivrus blocking outgoing traffic from the browser to the OpsCenter. – Gerard Mar 13 '15 at 10:43

2 Answers2

3

In certain environments the persistent connection between the browser and opscenterd may fail. We're working on implementing a more robust connection that will work in all environments, but in the meantime you can use the following workaround:

http://www.datastax.com/documentation/opscenter/5.1/opsc/troubleshooting/opscTroubleshootingZeroNodes.html

mbulman
  • 470
  • 2
  • 7
1

Minimal configuration that I find working was setting this options below for address.yaml

stomp_interface: [opscenter-ip]
stomp_port: 61620
use_ssl: 0
cassandra_conf: /etc/cassandra/cassandra.yaml
jmx_host: [cassandra-node-ip]
jmx_port: 7199

Make sure you have sysstat installed also.

Superpaul
  • 71
  • 5