* adding Operational Considerations section
modified: docs/getting-started-guides/ubuntu/logging.md
* Moving logging section from operational-considerations.md to logging.md
renamed: docs/getting-started-guides/ubuntu/operations.md -> docs/getting-started-guides/ubuntu/operational-considerations.md
* mapping to the proposed name
* ditched CDK naming in favor of k8s or cluster
* add whitelines for titles
Add documentation for Getting Started guides.
Kargo is an incubated Kubernetes project
https://github.com/kubernetes-incubator/kargo.
Signed-off-by: Bogdan Dobrelya <bogdando@mail.ru>
This page previously included instructions to use curl to download and install
Kubectl via http, rather than with https. This exposed the end user to
man-in-the-middle attacks.
Additionally, the instructions given here are slightly different from the
authoritative page at https://kubernetes.io/docs/getting-started-guides/kubectl/
- which covers useful things like getting the latest release and setting up tab
autocompletion.
Rather than just change 'http' to 'https' in the URLs here, I've updated the
minikube setup page to include a reference to the kubectl setup page.
- This supercedes PR #1770
- This breaks up the section from one huge installation page into:
- A new index page.
- New support grid with commercial and community support options.
- Move to lifecycle-based pages based on operational tasks.
- Add support for local deployments via LXD.
- Added screenshots to images/docs/ubuntu
- Add backup page.
- Add decommissioning page.
- Add a glossary of terms page.
- Rewritten installation page.
- Add logging page.
- Add monitoring page.
- Add networking page, flannel only for now, calico in progress.
- Add a scaling page.
- Add a security page.
- Add a storage page.
- Add a troubleshooting page.
- Add an upgrade page.
- Add a cluster validation page.
- Add new ubuntu content to the index page.
- Add Ubuntu to _data TOC.
- Add warning about choosing ipv6 with LXD, thanks Ed Baluf.
- Template-ize all the pages per jaredb's review.
* Example of `/etc/sysconfig/flanneld` is not valid
Configuration keys were different, I got those errors when followed steps from docs:
```
[ailyin@elite ~]$ systemctl status flanneld.service
● flanneld.service - Flanneld overlay address etcd agent
Loaded: loaded (/usr/lib/systemd/system/flanneld.service; enabled; vendor preset: disabled)
Active: activating (start) since Wed 2017-01-04 19:00:21 EST; 34s ago
Main PID: 4668 (flanneld)
Memory: 3.5M
CGroup: /system.slice/flanneld.service
└─4668 /usr/bin/flanneld -etcd-endpoints= -etcd-prefix=
Jan 04 19:00:46 elite.dev flanneld[4668]: E0104 19:00:46.395562 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:47 elite.dev flanneld[4668]: E0104 19:00:47.395788 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:48 elite.dev flanneld[4668]: E0104 19:00:48.395990 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:49 elite.dev flanneld[4668]: E0104 19:00:49.396141 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:50 elite.dev flanneld[4668]: E0104 19:00:50.396306 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:51 elite.dev flanneld[4668]: E0104 19:00:51.396451 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:52 elite.dev flanneld[4668]: E0104 19:00:52.396572 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:53 elite.dev flanneld[4668]: E0104 19:00:53.396701 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:54 elite.dev flanneld[4668]: E0104 19:00:54.396855 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
Jan 04 19:00:55 elite.dev flanneld[4668]: E0104 19:00:55.396992 04668 network.go:53] Failed to retrieve network config: client: etcd cluster is unavailable or misconfigured
```
* Updated Flanneld configuration example for node (same as for master);
Signed-off-by: yanan Lee <energylyn@zju.edu.cn>
Incorrect spelling
Signed-off-by: yanan Lee <energylyn@zju.edu.cn>
spelling error
Signed-off-by: yanan Lee <energylyn@zju.edu.cn>
Incorrect spelling
Signed-off-by: yanan Lee <energylyn@zju.edu.cn>
Revert "Incorrect spelling"
fix some typos
Signed-off-by: Jie Luo <luo612@zju.edu.cn>
fix a typo
Signed-off-by: Jie Luo <luo612@zju.edu.cn>
fix a typo
Signed-off-by: Jie Luo <luo612@zju.edu.cn>
Provide valid --cloud-providers, vagrant and openshift were listed but is not actually a valid cloud provider in this context in the current code base while a number of other valid options were omitted including azure, cloudstack, openstack, photon and vsphere.
The original one is: "kops update cluster useast1.dev.awsdata.com --yes"
However, the previous example used useast1.dev.example.com to create configuration file.
So, the "awsdata" should be changed into example.
Although the Fedora install docs are under the "Bare Metal"
section in the docs, the instructions provided work fine
for Fedora installs in virtual machines
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The Fedora multi-node doc need to tell users to install
flannel, since previous setup won't have brought it in.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The install of kubernetes pulls in docker, which has a direct
dependancy on iptables. So there is no need to tell people to
yum install iptables manually.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The /var/run/kubernetes directory is already created by
installation of the kubernetes RPM with correct user
and group ownership and permissions.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The docs still refer to etcd using port 4001, but in all
currently released Fedora versions it will be using 2379
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The KUBE_LOGTOSTDERR, KUBE_LOG_LEVEL and KUBE_ALLOW_PRIV
settings will already have correct values out of the box,
so no need to tell people to set them.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
There is no reason to tell people to enable updates-testing
repo for Fedora by default. The latest stable packages are
already in the main repos and updates-testing should only be
used by people wishing to test & report quality of pre-released
software, not general Kubnetes users.
Also remove the link to Koji builds - any Koji build which
is working well enough to user will be submitted as an update.
Telling users to install packages straight from Koji will
lead to them using potentially unstable / buggy builds.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
dnf has replaced yum as the standard package install tool
in all versions of Fedora that are currently supported.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
divyenpatel
Jorge O. Castro
Michael Mrowetz
Zhou Tao
Andrew Chen
Robert Roland
Andrew Martin
caiyixiang
Philipp T
Shawn McCarthy
Sean M. Collins
Steve Perry
Jared
Jeff Mendoza
Samuel Cozannet
Samuel Cozannet
Bogdan Dobrelya
luks
tolleiv
Matt Rickard
Oskar Pearson
Peter Lee
Francesc Rosas
bruceauyeung
Charles Butler
Devin Donnelly
Alexander Ilyin
Daniel Smith
gavinB-orange
Mik Vyatskov
Ritesh H Shukla
Alexander Kanevskiy
gbzhu
Aaron.L.Xu
dongziming
yanan Lee
tim-zju
SRaddict
Steve Gordon
king-julien
Matt Baldwin
Jitendra Bhurat
Jie Luo
Bilgin Ibryam
Casper CY Chiang
Brandon Philips
Ben Balter
king-julien
Brendan Burns
Joe Beda
Lucas Käldström
Cole Mickens
Janet Kuo
Martin Rauscher
Alejandro Escobar
Ahmet Alp Balkan
Saad Ali
Saksham Sharma
Daniel P. Berrange
Clayton Coleman
Jitendra Bhurat
Bryan Boreham