OpenShift Pod Operations: Difference between revisions

From NovaOrdis Knowledge Base
Jump to navigation Jump to search
 
(20 intermediate revisions by the same user not shown)
Line 6: Line 6:


* [[OpenShift_Operations#Subjects|OpenShift Operations]]
* [[OpenShift_Operations#Subjects|OpenShift Operations]]
* [[OpenShift Application Operations#Overview|Application Operations]]
* [[OpenShift Pod Concepts#Overview|OpenShift Pod Concepts]]


=List Pods for a Project=
=List Pods for a Project=
Line 26: Line 28:
  oc [-n ''<project-name>''] [[Oc_describe#pod|describe]] pod/''<pod-name>''
  oc [-n ''<project-name>''] [[Oc_describe#pod|describe]] pod/''<pod-name>''
  oc [-n ''<project-name>''] [[Oc_edit#Generic_Syntax|edit]] pod/''<pod-name>''
  oc [-n ''<project-name>''] [[Oc_edit#Generic_Syntax|edit]] pod/''<pod-name>''
=Monitor Pod's Logs=
[[oc logs]] -f <''pod-name''>
=Execute Commands in Containers=
[[oc rsh]] <''pod-id''> <''command''>
=Deploy a Simple Pod that Allows Arbitrary Commands in a Project=
{{Internal|OpenShift Simple Pod Running inside an OpenShift Project|Simple Pod Running inside an OpenShift Project}}


=Update a Pod Defintion=
=Update a Pod Defintion=
Line 37: Line 51:
The controller (either [[OpenShift_Concepts#Replication_Controller|replication]] or [[OpenShift_Concepts#DaemonSet|DamonSet]]) will detect that the pod is missing and it will start a new one.
The controller (either [[OpenShift_Concepts#Replication_Controller|replication]] or [[OpenShift_Concepts#DaemonSet|DamonSet]]) will detect that the pod is missing and it will start a new one.


=Scaling Down and Up the Number of Replicas associated with a DeploymentConfig=
=Evacuate Pods from Node=
 
{{External|https://docs.openshift.com/container-platform/latest/admin_guide/manage_nodes.html#evacuating-pods-on-nodes}}


This procedure can be applied after a configuration change, such as a modification in the [[OpenShift Concepts#ConfigMap|ConfigMap]].
Only pods backed by [[OpenShift Concepts#Replication_Controller|replication controllers]] can be evacuated. The [[OpenShift_Concepts#Bare_Pod|"bare" pods]] cannot be evacuated.


Identify the name of the deployment config:
Drain the node.


  oc get dc -o name --selector=....
  [[oadm drain]] <''node-name''> [--delete-local-data] [--ignore-daemonsets] [--force]


or
oadm drain node2 --delete-local-data --ignore-daemonsets


oc get dc
Use --delete-local-data if there are pods with local storage - such as Maven build pods.


Then set the number of replicas to 0:
Use --ignore-daemonset if there are ignore DaemonSets:  If there are DaemonSet-managed pods, drain will not proceed without --ignore-daemonsets, and regardless it will not delete any DaemonSet-managed pods, because those pods would be immediately replaced by the DaemonSet controller, which ignores unschedulable markings.


oc scale deploymentconfig/<''deploymentconfig-name''> --replicas=0
oadm manage-node --evacuate is deprecated.


After all pods associated with deployment configuration have been stopped, scale it back up again.
=Scaling Down and Up the Number of Replicas associated with a DeploymentConfig=


oc scale deploymentconfig/<''deploymentconfig-name''> --replicas=1
{{Internal|OpenShift_Application_Operations#Scaling_Up_and_Down_the_Number_of_Replicas_associated_with_a_DeploymentConfig|Application Operations - Scaling Up and Down the Number of Replicas associated with a DeploymentConfig}}

Latest revision as of 01:22, 28 February 2018

External

Internal

List Pods for a Project

oc get projects

oc get -n <project-name> pods

List Pods on Nodes

oadm manage-node <node-name> --list-pods

Pod Status

The most complete information set is generated by "oc get -o yaml pod ..." and "oc describe pod ...": 

oc [-n <project-name>] [[oc get#pods|get pods]
oc [-n <project-name>] -o yaml get pod] <pod-name>
oc [-n <project-name>] describe pod <pod-name>
oc [-n <project-name>] describe pod/<pod-name>
oc [-n <project-name>] edit pod/<pod-name>

Monitor Pod's Logs

oc logs -f <pod-name>

Execute Commands in Containers

oc rsh <pod-id> <command>

Deploy a Simple Pod that Allows Arbitrary Commands in a Project

Simple Pod Running inside an OpenShift Project

Update a Pod Defintion

Stopping Pods

A pod can be stopped by deleting its associated "pod" object: 

oc delete pod/logging-fluentd-3kz30

The controller (either replication or DamonSet) will detect that the pod is missing and it will start a new one.

Evacuate Pods from Node

https://docs.openshift.com/container-platform/latest/admin_guide/manage_nodes.html#evacuating-pods-on-nodes

Only pods backed by replication controllers can be evacuated. The "bare" pods cannot be evacuated.

Drain the node. 

oadm drain <node-name> [--delete-local-data] [--ignore-daemonsets] [--force] 

oadm drain node2 --delete-local-data --ignore-daemonsets

Use --delete-local-data if there are pods with local storage - such as Maven build pods.

Use --ignore-daemonset if there are ignore DaemonSets: If there are DaemonSet-managed pods, drain will not proceed without --ignore-daemonsets, and regardless it will not delete any DaemonSet-managed pods, because those pods would be immediately replaced by the DaemonSet controller, which ignores unschedulable markings.

oadm manage-node --evacuate is deprecated.

Scaling Down and Up the Number of Replicas associated with a DeploymentConfig

Application Operations - Scaling Up and Down the Number of Replicas associated with a DeploymentConfig
Retrieved from "https://kb.novaordis.com/index.php?title=OpenShift_Pod_Operations&oldid=38574"