[
https://issues.apache.org/jira/browse/METRON-898?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15986683#comment-15986683
]
Nick Allen commented on METRON-898:
-----------------------------------
Another setting, the `index_number_of_replicas` also needed to be changed from
2 to 1 in my environment. This might also be able to be calculated by the
Ambari MPack based on the number of data nodes that exist.
> Elasticsearch Starts in Red Status when Deployed with Ambari MPack
> ------------------------------------------------------------------
>
> Key: METRON-898
> URL: https://issues.apache.org/jira/browse/METRON-898
> Project: Metron
> Issue Type: Bug
> Reporter: Nick Allen
> Priority: Minor
>
> I deployed Elasticsearch using the Ambari MPack in an environment with 1
> master node (master only, no data) and 2 data nodes running on 3 separate
> hosts. Elasticsearch was able to start via Ambari, but the status was Red
> and it did not accept input or allow Kibana to query it.
> {code}
> GET _cluster/health
> {
> "cluster_name": "metron",
> "status": "red",
> "timed_out": false,
> "number_of_nodes": 3,
> "number_of_data_nodes": 2,
> "active_primary_shards": 0,
> "active_shards": 0,
> "relocating_shards": 0,
> "initializing_shards": 0,
> "unassigned_shards": 0,
> "delayed_unassigned_shards": 0,
> "number_of_pending_tasks": 0,
> "number_of_in_flight_fetch": 0,
> "task_max_waiting_in_queue_millis": 0,
> "active_shards_percent_as_number": "NaN"
> }
> {code}
> {code}
> GET _cluster/nodes/stats
> {
> "error": {
> "root_cause": [
> {
> "type": "cluster_block_exception",
> "reason": "blocked by: [SERVICE_UNAVAILABLE/1/state not recovered
> / initialized];"
> }
> ],
> "type": "cluster_block_exception",
> "reason": "blocked by: [SERVICE_UNAVAILABLE/1/state not recovered /
> initialized];"
> },
> "status": 503
> }
> {code}
> I was able to fix this manually by changing the
> `gateway_recover_after_data_nodes` from 3 to 2. Since I only have 2 data
> nodes, this setting cannot be 3. It should not be greater than the number of
> data nodes.
> I would expect the Ambari MPack to be able to determine a working setting for
> this parameter based on the number of data nodes, so that Elasticsearch works
> out-of-the-box.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
