Plugin is always disabled

[kesyk]

I am deploying opendistro elasticsearch as single node. And since last week i cannot make this plugin work(before it had worked and i did not change something). In logs i see:

[2020-09-07T11:21:24,156][INFO ][c.a.o.e.p.h.c.PerformanceAnalyzerConfigAction] [a7d349060c8c] PerformanceAnalyzer Enabled: false

I have tried different PCs with fresh docker, deploying by single run command and docker-compose files that are in official documentation. I POSTed to localhost:9200/_opendistro/_performanceanalyzer/config and /_opendistro/_performanceanalyzer/cluster/config with {"enabled": true} or {"pa_enabled": true}. Anyway it does not work. Every time i redeploy it, i clear everything from docker with docker system prune -a --volumes Es version: 1.9.0.

[Franckiboy15]

I have the same issue, any news on what prevent the plugin to start?

[kesyk]

Previous version made it start(

On Mon, Nov 30, 2020, 21:47 Franckiboy15 [email protected] wrote:

I have the same issue, any news on what prevent the plugin to start?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/opendistro-for-elasticsearch/performance-analyzer/issues/199#issuecomment-736002385, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJDZFVXBSON46FYGDSCXVCTSSPZG5ANCNFSM4Q6BY7PA .

[Franckiboy15]

I just update from 1.10.1 to 1.11.0 and I got the same issue I know that previously it was working on 1.9.0 since I got the issue that the Index was growing indefinitely. As it been disable since then in the docker version?

[sruti1312]

During startup, cluster API to disable PA does not work. We are working on fixing this (https://github.com/opendistro-for-elasticsearch/performance-analyzer/issues/232)

This is a problem only during startup. After re-enabling PA (cluster_state=3), the disable option for PA works fine.

[ec2-user@ip-10-212-47-103 logs]$ curl localhost:9200/_opendistro/_performanceanalyzer/cluster/config -H 'Content-Type: application/json' -d '{"enabled": false}'
{"currentPerformanceAnalyzerClusterState":3,"shardsPerCollection":0,"batchMetricsRetentionPeriodMinutes":7}
[ec2-user@ip-10-212-47-103 logs]$ curl -XGET localhost:9200/_opendistro/_performanceanalyzer/cluster/config
{"currentPerformanceAnalyzerClusterState":3,"shardsPerCollection":0,"batchMetricsRetentionPeriodMinutes":7}

From logs:

[2020-10-30T16:05:08,008][INFO ][o.e.c.s.ClusterSettings  ] [c4d4a248a38f4230f583d5cd1cf475ee] updating [cluster.metadata.perf_analyzer.state] from [] to [0]

[ec2-user@ip-10-212-47-103 logs]$ curl -s 'localhost:9200/_cluster/settings?pretty&include_defaults&flat_settings' | grep perf
  "cluster.metadata.perf_analyzer.state" : "0",
  "cluster.metadata.perf_analyzer.config.overrides" : "",
  "cluster.metadata.perf_analyzer.pa_node_stats_setting" : "1",

Intermittent solution:

1. Disabling RCA followed by disabling PA works.

curl localhost:9200/_opendistro/_performanceanalyzer/rca/cluster/config -H 'Content-Type: application/json' -d '{"enabled":false}'
curl localhost:9200/_opendistro/_performanceanalyzer/cluster/config -H 'Content-Type: application/json' -d '{"enabled": false}'

2. Use Node API's.