Is there a limit to the number of nodes displayed in the free version of Marvel?
Right now I have 12 data nodes, all showing up, but I have 12 logstash nodes as well with ES client nodes running on them. When I view Marvel I only see a few of the logstash nodes. The configs are identical because I copied them from one node to all others. I know the nodes are there because:
curl 10.1.55.2:/_cat_nodes | grep INDEXER
10.1.55.6 10.1.55.6 4 9 0.78 - - INDEXER_NODE_4
10.1.55.27 10.1.55.27 1 9 0.88 - - INDEXER_NODE_9
10.1.55.30 10.1.55.30 0 9 0.59 - - INDEXER_NODE_12
10.1.55.25 10.1.55.25 1 9 0.56 - - INDEXER_NODE_7
10.1.55.26 10.1.55.26 2 9 0.35 - - INDEXER_NODE_8
10.1.55.24 10.1.55.24 3 9 0.81 - - INDEXER_NODE_6
10.1.55.28 10.1.55.28 2 9 0.48 - - INDEXER_NODE_10
10.1.55.4 10.1.55.4 0 9 0.68 - - INDEXER_NODE_2
10.1.55.23 10.1.55.23 3 9 0.29 - - INDEXER_NODE_5
10.1.55.29 10.1.55.29 1 9 0.31 - - INDEXER_NODE_11
10.1.55.5 10.1.55.5 3 9 0.66 - - INDEXER_NODE_3
This shows all the indexer nodes I have running right now. (I know node 1 is not listed it is currently offline.)
Why would all of the nodes be listed when running the curl command but not show up in Marvel.
I have confirmed I am running the same version of elasticsearch, logstash, and marvel-agent on all nodes
ES 2.3.1
LS 2.3.1
Marvel-Agent 2.3.1
License 2.3.1
I have not seen this issue and there's no realistic limit that is set.
On the cluster being monitored, can you run:
$ curl -XGET host:9200/_cat/plugins?v
This will show what nodes believe to be installed as opposed to what's on the file system. It should list 12 data nodes, 12 client nodes (your "indexers"), and presumably 3 master nodes. Each one should show marvel-agent installed for the same version.
Then just attach the output either here or as a gist (posting it as a gist makes it easier for you to delete later).
At the same time, can you also add any configuration settings that you are applying to one of the nodes that is missing and one of the same types that is not missing?
So I'm not sure what happened but after 2 days of not seeing some of the nodes in marvel now I am seeing all of them. Its strange. I made no configuration changes at all.
I'm going to chalk it up to having to wait for some reason.
Thanks for the help. I am still happy to provide the requested info if it will help you troubleshoot this issue. But I am considering it resolved.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.