| Summary: | [DOCS] Lack of information on NodePorts | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Eric Jones <erjones> |
| Component: | Documentation | Assignee: | Gaurav Nelson <gnelson> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | zhou ying <yinzhou> |
| Severity: | medium | Docs Contact: | Vikram Goyal <vigoyal> |
| Priority: | high | ||
| Version: | 3.1.0 | CC: | aos-bugs, ccoleman, erich, erjones, jkaur, jokerman, knakayam, mmccomas, nicholas_schuetz, pcameron |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | 3.7-release-plan | ||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2017-10-26 01:11:36 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Eric Jones
2016-03-02 15:22:46 UTC
Bug number 1303232 was closed in favor of 1287471 They both make some reference to using nodePorts to achieve the goal of using non-HTTP(s) routing and they both suggest that the documentation should exist on how to do this. @Clayton, as you made these suggestions, can you help create this documentation? Updated the docs to include the example provided in attached solution. Please review the changes in PR https://github.com/openshift/openshift-docs/pull/5237 Hello Eric, can you please review this PR? erjones I am not sure what you would like to achieve with HA and node ports. The nodeport is bound to the service on every node in the cluster. So if you can get traffic to any node in the cluster it will go the service which will load balance it to some endpoint. See: https://docs.openshift.com/container-platform/3.6/dev_guide/getting_traffic_into_cluster.html This should have what you need. If not add the info there. HA is different for routers and services. Routers run on specific nodes in the cluster and external network admins must get the desired traffic to the node that is running the router. To make this high availability we increase the router replicas and use ipfailover to make a virtual IP (VIP) land on one of the nodes that is running a router replica. The external admin must get traffic to the VIP. The VIP will be on a node with a router. For services, the nodeport is listening on all nodes in the cluster. The service will have multiple endpoints (typically the DC or RC has replicas > 1) If you manage to get traffic to any of the nodes in the cluster the nodeport will connect and traffic will go to one of the endpoints. The V1 docs (referenced above) are out of date. I would expect NodePort to scale at least as high as anything except dedicated routing and switching technology or hardware assisted routing, but you'll need to have a layer in front to ensure traffic reaches the appropriate nodes (whether that's a TCP load balancer, ECMP routing, etc). I commented on the Pr Thanks Clayton, I have updated the PR to include information provided. Gaurav Nelson: The PR LGTM, will verify this issue. Commit pushed to master at https://github.com/openshift/openshift-docs https://github.com/openshift/openshift-docs/commit/e1cff0c7001bc328d7aa6e87d315668666ff5d8f Merge pull request #5237 from gaurav-nelson/Bug1313910Fixes Bug#1313910 fixes - added information on nodePorts |