rhsyseng / docker-rhel-elasticsearch Goto Github PK

Is there any API which can be called before shutting down the node? For example Cassandra uses nodetool drain

@lukas-vlcek do you use this in ECL?

ES fails to start if vm.max_map_count != 262144 https://www.elastic.co/guide/en/elasticsearch/reference/current/vm-max-map-count.html.

ECL uses ansible to set this https://github.com/openshift/openshift-ansible/blob/master/playbooks/openshift-logging/private/config.yml#L37

See: https://github.com/pires/kubernetes-elasticsearch-cluster/blob/master/es-client.yaml#L22

#37 introduced new Openshift deployment file. At the moment there are two deployment file with some repetition. I am wondering whether we could split them so some parts could be reused. For example service definitions are always the same.

#44 bumped ES version from 5.5.2 to 5.6.10 but it's not available in docker registry.

When does https://github.com/CentOS/container-index rebuild the images?

Shouldn't be this line
https://github.com/CentOS/container-index/blob/master/index.d/rhsyseng.yaml#L98 changed to 5.6.10?

cc @ruromero

I needed to sudo sysctl -w vm.max_map_count=262144 on the OpenShift nodes running the pods, as elastic bootstrap tests failed otherwise...

A common label can be used to manipulate (e.g. delete) all objects defined in this template

Remove service-account it's not needed. It was used by elasticsearch-cloud-kubernetes which has been removed see #9

e.g. https://github.com/RHsyseng/docker-rhel-elasticsearch/blob/master/Dockerfile#L31

Can we remove https://github.com/fabric8io/elasticsearch-cloud-kubernetes plugin for unicast discovery?

For example https://github.com/pires/kubernetes-elasticsearch-cluster is using Zen discovery https://www.elastic.co/guide/en/elasticsearch/reference/5.6/modules-discovery-zen.html

There is https://github.com/pires/kubernetes-elasticsearch-cluster/blob/master/es-discovery-svc.yaml which points to master nodes (there is no es-master service). Then all nodes define discovery.zen.ping.unicast.hosts pointing to this service.

Some links:

• https://github.com/pires/kubernetes-elasticsearch-cluster
• https://github.com/pires/docker-elasticsearch-kubernetes
• https://github.com/pires/docker-elasticsearch

By default template deploys two master nodes, therefore NODE_QUORUM should be set to 2. Currently it is 1 which results in log:

[2018-04-16T09:51:50,061][WARN ][o.e.d.z.ElectMasterService] [elasticsearch-0] value for setting "discovery.zen.minimum_master_nodes" is too low. This can result in data loss! Please set it to at least a quorum of master-eligible nodes (current value: [1], total number of master-eligible nodes used for publishing in this round: [2])

https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-node.html#split-brain

Elasticsearch should have readiness probe.

Some references:

• https://trello.com/c/H1kkjLQU/437-3-elasticsearch-should-have-readiness-check
• openshift/origin-aggregated-logging#358
• fabric8io/elasticsearch-cloud-kubernetes#90
• https://github.com/pires/kubernetes-elasticsearch-cluster/blob/master/es-client.yaml#L64 - only client nodes are using it

Port the following PR to master:
#37 #36 #35

Openshift resource files could be easily used with different ES image (e.g. configuration is in configmap).

What would need to change? Use openshift template?

For example ECL is setting half of the available memory as heap size https://github.com/openshift/origin-aggregated-logging/blob/master/elasticsearch/run.sh#L76

Hi,

I am trying to deploy this on minishift minishift v1.12.0+daa0943, openshift v3.6.1+008f2d5 and I get the following error. See the last event:

Name:			elasticsearch
Namespace:		myproject
CreationTimestamp:	Thu, 15 Feb 2018 15:44:27 +0100
Selector:		app=elasticsearch
Labels:			app=elasticsearch
Annotations:		image.openshift.io/triggers=[{
"from": {
  "kind": "ImageStreamTag",
  "name": "elasticsearch:6.2.1"
  },
  "fieldPath": "spec.template.spec.containers[?(@.name==\"elasticsearch\")].image"
}]

Replicas:	2 desired | 0 total
Pods Status:	0 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:		app=elasticsearch
  Service Account:	elasticsearch
  Containers:
   elasticsearch:
    Image:	 
    Ports:	9200/TCP, 9300/TCP
    Limits:
      memory:	512Mi
    Requests:
      memory:	512Mi
    Environment:
      SERVICE:		elasticsearch-cluster
      LOG_LEVEL:	info
    Mounts:
      /elasticsearch/persistent from elasticsearch-persistent (rw)
  Volumes:
   elasticsearch-persistent:
    Type:	EmptyDir (a temporary directory that shares a pod's lifetime)
    Medium:	
Volume Claims:	<none>
Events:
  FirstSeen	LastSeen	Count	From		SubObjectPath	Type		Reason		Message
  ---------	--------	-----	----		-------------	--------	------		-------
  50s		12s		15	statefulset			Warning		FailedCreate	create Pod elasticsearch-0 in StatefulSet elasticsearch failed error: Pod "elasticsearch-0" is invalid: spec.containers[0].image: Invalid value: " ": must not have leading or trailing whitespace

On fedora 28 docker docker run --rm -it --net=host registry.centos.org/rhsyseng/elasticsearch:5.5.2 freezes and fails.

sysctl vm.max_map_count returns vm.max_map_count = 262144