After deploying the PV and PVC via the YAML file, run `kubectl describe pvc mysql-pv-claim` should show `mysql-pv-volume` as the `Volume`, instead of `mysql-pv`.
This heading is missing in the 1.10 documentations which makes it hard to find the needed information because the anchor `setting up a master load balancer` points to nowhere. This was working in the 1.19 documentation and it was really useful.
I noticed that the three last headers were missing from the index, so I inspected the markdown and realized that they had only one hashtag (#) as compared to two (##) as all the other headers had.
I simply made those three topics follow the same pattern I observed the other did.
# Issue
The command of jobs does not run correctly if its job already finished.
Some readers may confuse that they can't succeed to run a command to follow this tutorial.
So, we should add `--show-all` option to show pods that may be completed.
# The log to show that we need to add an --show-all option
```
$ kubectl get jobs
NAME DESIRED SUCCESSFUL AGE
busybox 1 0 27d
hello-1529501040 1 1 2m
hello-1529501100 1 1 1m
hello-1529501160 1 1 43s
$ kubectl get pods --selector=job-name=hello-1529501160
No resources found, use --show-all to see completed objects.
$ kubectl get pods --show-all --selector=job-name=hello-1529501160
NAME READY STATUS RESTARTS AGE
hello-1529501160-dckjv 0/1 Completed 0 1m
```
Those links are dead, pointing at what I assume in the previous place for the Kubelet TLS bootstrapping documentation.
Update them with the new location!
* Update etcd config description of kubeadm HA doc
- Fix url from http to https
- Add etcd-listen-ip description
* Edit for clarity and list markup
* More minor edits
implementation-details.md is outdated for k8s 1.11.
Add feature state that this page applies to k8s 1.10.
The page should be updated for k8s 1.11 in a later stage
in the release cycle.
`base64` generates 76-character chunks of text, delimited by a space character. the `=+/` characters are removed, so on the (uniformly random) offchance that 44 characters are removed, the last `dd` will read a whitespace, and the token will include a whitespace, which "usually" isn't defined as an "alphanumeric character." Removing the space character as well, would mean that 142 of the characters would have to be a `=+/[:space:]`, before the token becomes shorter.
Cannot copy any apiserver* keys and certs, as they are generated on startup. If not, it results in an error from kubeadm
```
certificate apiserver is not signed by corresponding CA
```
Also, if the kubernetes.list does not exist you will not be able to edit it, so adding touch and then add the line will prevent errors with installing via apt-get.
It is confusing to capitalize "container" because it isn't really a primitive or API object. Furthermore, this guideline doesn't appear to be closely followed.
* Overall cleanup for kubeadm setup guide.
* Add sonobuoy verification next step
* Partial copyedits
Hit a place where we also have formatting issues, plus a TODO (removed bc these changes go live as soon as merged). More work to do, but this should help the bottleneck for now.
* Address comments from reviewers
* Copy edits, fiddly bits, and clearer warnings
Incorporated feedback from @mistyhacks, fixed many nits, reorganized some **Note:** content more sensibly.
* Revised title per feedback
* Add a note regarding escaping characters
Updated docs with a note regarding escaping characters, it's important because it won't work well otherwise.
* Update secret.md
* No escape required for --from-file
Feedback from @mengqiy
Added information to install cfssl and cfssljson on all etc nodes as they are required to generate clients certificates
Change <podname> to <name> in static pod yaml description for etcd deployment
Additionaly the first bash script in "Run kubeadm init on master0" section doesn't render correctly on kubernetes.io site (https://kubernetes.io/docs/setup/independent/high-availability/) but I don't see any pb with markdown.
* Issue #8493 - Missing instructions in "Creating HA clusters with kubeadm":
- change to kubeadm-init.yaml: configure `advertiseAddress` to the VIP
- added section with instructions for configuring `kubelet`, `kube-controller-manager` and `kube-scheduler` to use the VIP
* Issue #8493 - Missing instructions in "Creating HA clusters with kubeadm": Change use the 'controlPlaneEndpoint' setting in kubeadm-init.yaml, hence no configuration file patching afterwards is necessary anymore.
* Issue #8493 - added private IP to apiServerCertSANs as proposed by @detiber
These pages did not use any template and the displayed toc started
with two bullet points.
Fixes:
* Use the concept template everywhere and add capture markers.
* Rework content-organization.md to move the references to a "whatsnext"
section.
* Fix in localization.md a misparsed list.
* Getting started gce deadlink
1.
modify for kubernetes/website link
2.
remove unnecessary .md link
* modify link from github.com to kubernetes.io
* revert blogs modification
* revert blog post for meta char
Fix some hugo transition issues:
* Use `file` as code syntax following our style guide and fixing
bug #8868
* Reformat Note so that it's not displayed as text.
Closes: #8868
In _index doc the Kubectl CLI about overview link
was add in "[Overview of kubectl]", So remove not
need "[kubectl]" link.
Signed-off-by: Yuanbin.Chen <cybing4@gmail.com>
* Fix errors in configure-multiple-schedulers.md
- Failed to push docker image
- Errors found in my-scheduler logs
```
...
E0515 06:36:28.180428 1 reflector.go:205] k8s.io/kubernetes/vendor/k8s.io/client-go/informers/factory.go:130: Failed to list *v1beta1.ReplicaSet: replicasets.extensions is forbidden: User "system:serviceaccount:kube-system:default" cannot list replicasets.extensions at the cluster scope
```
* Replace double quote with backticks for code strings
* Use `system:kube-scheduler` instead of `cluster-admin`
* Create dedicated service account for `my-scheduler`
* Update high-availability.md
[fix]command should leading
* Repair code inside list items, remove "like so"
* Modify one code block, change 1-step procedures to not be procedures
This commit adds details on how to create Windows services from
kubelet and kube-proxy.
Signed-off-by: Alin Balutoiu <abalutoiu@cloudbasesolutions.com>
* Add link: add changelog-1.10 and changelog-1.11
* Delete download page's catalog: hugo's markdown render engine can generate catalog through the TOC
* Format the release list
* Delete the release note
In recent versions of Kubernetes, selector field in DeploymentSpec is required and it'll not be filled with values from pod template. simple_deployment.yaml was already corrected and selector was added, here in this commit the content of the docs corrected to match this change.
* Clarify that beta annotations can not be removed
Removing annotations caused a major outage with nodes that lagged the
master. We used annotations as would-be fields, though we don't any
more. Now we just add fields. That makes them part of the API version.
These annotations are effectively part of the API version, too.
It stinks because I love to delete code, but I think it's just part of
the debt-accrual process and yet another reason we need to be judicious
in adding fields.
* Change 6 month deprecations to 9 months
This covers maximum version skew. For example a master might be at
1.10, while nodes are at 1.8. We must not break compat within that
window.
* Update service-access-application-cluster.md
Fix code snippet formatting and indentation in ordered list (#8698)
* Fixes code formatting using code fences
Here's a pass with code fences. It seems to work well, and I indented the code fences in the source for readability, even though it wasn't needed for proper indentation in the rendered output.
I wasn't able to use code fences and get the text "The response to a successful request is a hello message:" on its own line. I can achieve that using indentation, but for consistency's sake I used the code fences and placed the text with the paragraph above it.
* Adds `shell` to the code fences
* Fix etcd installation command in kubeadm tutorial
Shells interpolate variables before running commands, so setting the
ETCD_VERSION variable must be separated from the command that
interpolates it. Otherwise an empty string will be interpolated.
The rm command isn't necessary - the output from curl is piped directly
to tar.
* Update stated default etcd version for K8s 1.10
1.10 is the most recent version of K8s and the default etcd server
version for it is 3.1.12
* Fix non-US English typo
Style guide states US English should be used
* Add configure link description
I added a configure link description. Just having a link
on a page with no introductory context seemed awkward.
Signed-off-by: Brad Topol <btopol@us.ibm.com>
* removing unnecessary header
* updated text to be active voice
A tab shortcode was using angle brackets even though the
contents were markdown. This caused a formatting issue.
This switches the tab to use the `%`-delimiter to properly
render the markdown.
This commit adds a missing closing capture and puts "## Install kubectl" inside the relevant capture.
Note that the missing close was also missing before Hugo-migration, so the section ordering is still not the same as the Jekyll version of the site. But I think it is more corret.
Note that content outside the capture template blocks, is included as defined is this template:
https://github.com/kubernetes/website/blob/master/layouts/partials/templates/task.html#L8
See #8277
* Fix url to etcd binary
etcd should be 3.12 for kubernetes 1.10, also fix the URL (probably was broken after migration to Hugo)
* Make Curl command safer
You really only need that variable to be set for the specific command, so no need to export it.
Also, when using Bash variables in a string, you should enclose them in curly braces to delineate them from the rest of the string. This helps prevent problems where Bash can't tell where the name of the variable ends and the next part of the string begins.
The page was wrongly formatted:
* Two lists were shown as paragraphs, add empty lines so that they are
parsed as lists.
* The list of members, reviewers was not intended properly, so that the
single items are intended but not their description. Indent
everything.
* Fix URL to owners.md file, the old one gives a 404.
* Fix code blocks in set-up-cluster-federation-kubefed
Two of the code blocks don't render correctly in the [docs page](https://kubernetes.io/docs/tasks/federation/set-up-cluster-federation-kubefed/) (see *Adding a cluster to a federation* section) even though they render correctly in the GitHub preview. Hopefully this PR will fix it.
Plus fixed a few instances where the code blocks in set-up-cluster-federation-kubefed.md don't indicate the code block's language.
* Fix spacing
- Use environment variables when redirecting to "/etc/etcd.env".
- Add more placeholder values for flags in "etcd.service" -> "ExecStart"
Fixeskubernetes/kubeadm#797
* Update kubeadm-init.md
Fix broken kubeadm token link
* Fix more broken links in kubeadm-init.md
Broken links are fixed including kubeadm-join, kubeadm-upgrade, and kubeadm-reset
Currently, the link just points to a description of
what Kubernetes objects are and does not
explicitly state what a declarative API is. It is
not immediately clear to the reader what the term
means.
This commit tries to explain it explicitly so that
the link adds to the meaning and it is clearer to
the reader.
ThirdPartyResources (TPRs) were removed in v1.8. It is misleading
to still show the docs for TPRs for > v1.8. If one needs to read
docs for TPRs, they can switch to the v1.8 version of the docs
website.
However, the docs of migration of TPRs to CRDs are still not
removed because one can also migrate from v1.7 (when CRDs were
introduced) to v1.7+.
First, fix https:// address.
Secondly, previous formatting would display on the page as one line
altogether, which couldn't be pasted into a terminal correctly.
The TOC looked like:
3. Deployments
1.1 Deployment Management
4. Services
1.1 Service Management
The problem is that the headings went from "##" to "####", use "###" for
second level heading to cleanup TOC.
fix some hugo issues:
* Add missing empty line for list.
* Indent items correctly to avoid overlong line shown as code.
* Fix broken table
Closes: #8457
Before running `kubeadm init` users should be informed that the command might change depending on the network plugins they choose. It is better to provide it before presenting with `kubeadm init` so that they don't have to tear the cluster apart once they figure out they needed some extra parameters while running the command
* Container Linux doc fixes
portions of the URLs were missing their bash substitutions and the variable exports don't apply to the following commands without using 'export'
* Adjustments from PR
export is not needed, used proper bash expansion
* typos
* Typo fix: formated->formatted and proper noun in capital
formated->formatted
Pod and Container are proper noun, in this doc, someplace are in capital, some are in lowercase. It's better to keep consistency.
* Update volumes.md
1) Remove missing links about kube-apiserver, kube-controller-manager
and kube-scheduler. Link to the high-availability page, which
isn't exactly a full replacement.
2) Add section about kube-proxy parameters.
Hugo does not support the way the image's max-width was specified before.
This fix will constrain images to always be 50% rather than max out
at 50%, but it is the best we can easily do for now.
Signed-off-by: Misty Stanley-Jones <mistyhacks@google.com>
* Add links to the F5 BIG-IP Controller for Kubernetes to the Ingress controllers section of the ingress doc
* Address review comments from @zacharysarah
* Fix concepts/api-extension/custom-resources.md
All tables were wrongly formatted, add proper second line so that hugo
formats them nicely again.
Add missing empty line before lists so that they get displayed as lists.
Closes: #8392
* Fix one more table
Fix another wrong table heading in custom-resources.md as noticed
during review.
Reading over the .md file, I fixed the following problems:
* Replace links to old location with new location.
* Fix missing shell quotation marks and shell annotations
* Fix wrong indent so that lists are shown correctly.
* Fix broken table, the "|" needs to be quoted to be rendered.
The way we previously did it in Jekyll doesn't work
with Black Friday. There are a few workarounds, but
the easiest way to fix it is to just change those
links to HTML email.
* [Issue #8336] Added a # symbol to What's next
Looking to see if the Table of contents will render properly if all the headers are H4 (####)
* use h2 instead of h4
* More readable fix for issue #7957
Replaces HTML entity-escaped text with Markdown-escaped text where necessary.
* Fixing the markdown discrepancy
Rendering consistently is for chumps