Errata

Encountered an error when trying out following code snippet

spec:
template:
annotations: kubernetes.io/ change-cause: "Update nginx to 1.9.10" ...


I think it should be

...
spec:
template:
metadata:
annotations: kubernetes.io/ change-cause: "Update nginx to 1.9.10"

Note from the Author or Editor:
The reporter is correct, 'metadata' is needed there. The indentation is important though, there should be 2 additional spaces on each subsequent line, it should look like this:

spec:
template:
metadata:
annotations: kubernetes.io/ change-cause: "Update nginx to 1.9.10"

Thanks

In the section, there is a sentence that says as follows:

> The ReplicaSet configuration file in Example 8-1 will ensure one copy of the gcr.io/kuar-demo/kuard-amd64:1 container is running at a given time.

But it seems the configuration file tries to run a copy of "gcr.io/kuar-demo/kuard-amd64:2".

Note from the Author or Editor:
Reporter is correct 'gcr.io/kuar-demo/kuard-amd64:1' should be 'gcr.io/kuar-demo/kuard-amd64:2'

(from Safari books online, dont know the page number)

Under Building Applications with Docker in Ch 2, the thrid line of teh Docker file should read:
COPY bin/1/amd64/kuard /kuard

This is actually fixed in the Readme for the repo, per Issue #7.
I list it here because it is incorrect in the book, and for those following along in the book, it can cause a hiccup.

Note from the Author or Editor:
Now fixed in the original source

This section states the default redis port is 6380, when in fact it is 6379.

Note from the Author or Editor:
Good catch. I'm changing the example to use port 6379.

Chapter 5. Pods > Pod Manifest > Creating a Pod Manifest

apiVersion: v1
kind: Pod
metadata:
name: influxdb
spec:
containers:
- image: b.gcr.io/kuar/influxdb:0.9.4.2
name: influxdb
ports:
- containerPort: 8083
name: admin
protocol: TCP
- containerPort: 8086
name: http
protocol: TCP

Error:
image can't be downloaded

Reason:
"On February 28th, 2017, the use of “bring-your-own-bucket” registries such as b.gcr.io and bucket.gcr.io is deprecated. Container Registry no longer serves any container images hosted in those buckets."
https://cloud.google.com/container-registry/docs/support/deprecation-notices

Into chapter 14: "Deploying Real-World Applications", onto the Ghost section, the authors suggest we create a ConfigMap from a file the following way:

$ kubectl apply cm --from-file ghost-config.js ghost-config

I tried this, and this caused an error on kubectl. I changed the command to:

$ kubectl create cm --from-file ghost-config.js ghost-config

The latter worked, as I believe the authors intended.


I hope this may help.

Kind Regards!
Alberto Biasao

Note from the Author or Editor:
The reporter is correct 'kubectl apply' should be 'kubectl create'

Thanks!

A project that is standardizing the container image format is not Open Container Image (OCI) project but Open Container Initiative (OCI) https://www.opencontainers.org/.

Note from the Author or Editor:
The reporter is correct. "Open Container Image" should be "Open Container Initiative"

Chapter 2 summary refers to Docker's ability to mount external directories, yet the chapter did not cover this topic!

Note from the Author or Editor:
The reporter is correct. I re-worked the summary in the books Atlas repository.

This command ...

$ gcloud container clusters create kuar-cluster


Does not produce this ...

$ kubectl get nodes
NAME         STATUS         AGE
kubernetes   Ready,master   45d
node-1       Ready          45d
node-2       Ready          45d
node-3       Ready          45d


But this ...

$ kubectl get nodes
NAME STATUS AGE VERSION
gke-kuar-cluster-default-pool-c7891302-2f1v Ready 42m v1.7.6
gke-kuar-cluster-default-pool-c7891302-34pk Ready 42m v1.7.6
gke-kuar-cluster-default-pool-c7891302-8q56 Ready 42m v1.7.6


The kubernetes master does not display, and the nodes are not named, "node-1", etc.

Note from the Author or Editor:
Edited the original. This is tricky because it will never match exactly. Added a note to this effect.

Text claims that running:

kubectl get daemonSets --namespace=kube-system kube-proxy

should show kube-proxy. It returns the below error. And, omitting kube-proxy on the command line only shows fluentd, not kube-proxy.

[tom@MacBook-Pro python]$ kubectl get daemonSets --namespace=kube-system kube-proxy
Error from server (NotFound): daemonsets.extensions "kube-proxy" not found
[tom@MacBook-Pro python]$ kubectl get daemonSets --namespace=kube-system
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE-SELECTOR AGE
fluentd-gcp-v2.0 3 3 3 3 3 beta.kubernetes.io/fluentd-ds-ready=true 20m

Note from the Author or Editor:
Good catch. Not all clusters will use DaemonSets to run the proxy. It can be configured and managed outside of Kubernetes. I'm updating the section to make it clear that many clusters will use a DaemonSet here but not all.

Two references to the UTC namespace. I believe this should be the UTS namespace.

Note from the Author or Editor:
Yup -- good catch. This has been updated in the latest versions of the book.

The command provided to show details of the restart is wrong: It should read "kubectl describe pod kuard" instead of "kubectl describe kuard".

Note from the Author or Editor:
The reporter is correct.

'kubectl describe kuard'

should be:

'kubectl describe pods kuard'

The "Mounting the host filesystem" chapter talks about "hostDir" volumes whereas the previous example referenced and the Kubernetes documentation talk about "hostPath" volumes.

Note from the Author or Editor:
Reporter is correct 'hostDir' should be 'hostPath' everywhere.

The following code example:

kubectl label pods bar -color

The dash must be after the "color"? E.g:

kubectl label pods bar color-

Note from the Author or Editor:
The reporter is correct, '-color' should be 'color-'

The respective section in the kuard UI is called "DNS Query", not "DNS Resolver".

Note from the Author or Editor:
Correct, please replace "DNS Resolver" with "DNS Query"

The respective kuard section is called "Readiness Probe", not "Readiness Check".

Note from the Author or Editor:
Fixed in repository.

The respective section in the kuard example application is called "Server Env", not "Environment". In the table shown, the first column is called "Key", not "Name".

Note from the Author or Editor:
Fixed in book repository.

"If you want to remove a label, you can use the -<label-name> syntax:

$ kubectl label pods bar -color


This should instead be "<label-name>- syntax":

$ kubectl label pods bar color-


The above correction was found from "kubectl label --help":

# Update pod 'foo' by removing a label named 'bar' if it exists.
# Does not require the --overwrite flag.
kubectl label pods foo bar-

Note from the Author or Editor:
The reporter is correct '-color' should be 'color-'

The sentence:

"You should see that the server is not returning 500s."

Must contains redundant "not"? E.g. is must read:

"You should see that the server is returning 500s."

Note from the Author or Editor:
Confirmed. 'not' should be deleted.

In the example of a couple application architecture in Figure 12-1, the boxes at the bottom are labeled as "Frontend". However, frontend servers are at the top so the ones at the bottom should be "Backend".

Note from the Author or Editor:
Reporter is correct. Boxes on the lower right-hand side of the illustration should be labelled "Backend #1" and "Backend #N", instead of "Frontend #1" and "Frontend #N".

The boxes on the upper right hand side of the illustration are labelled correctly.

The formula for estimating the amount of time for the rollout should refer to `minReadySeconds` instead of `maxReadySeconds`

Note from the Author or Editor:
Confirmed and fixed in source.

In the replicaSet Chapter on page 135 there is a command to imperatively scale up as follows ` kubectl scale kuard --replicas=4` that returns an error in the console.

The Error reads: error: resource(s) were provided, but no name, label selector, or --all flag specified

I was able to get the command to execute by adding `rs` to the command as follows.
`kubectl scale rs kuard --replicas=4` which then returned this message in the console "replicaset "kuard" scaled".

Note from the Author or Editor:
The reporter is correct. The command should read:

kubectl scale replicasets kuard --replicas=4

Thanks!

The paragraph reads:

The key names for data items inside of a secret or ConfigMap are defined to map to valid environment variable names. They may begin with a dot followed by a letter or number. Following characters include dots, dashes, and underscores. Dots cannot be repeated and dots and underscores or dashes cannot be adjacent to each other. More formally, this means that they must conform to the regular expression [.]?[a-zAZ0-9]([.]?[-_a-zA-Z0-9]*[a-zA-Z0-9])*.

As I understand the intent of the paragraph the name `.a1e.---_beta` should be invalid, but it is a match against the supplied regex pattern.

Note from the Author or Editor:
Confirmed. The correct regular expression is:

^[.]?[a-zAZ0-9]([.]?[a-zA-Z0-9]+[-_a-zA-Z0-9]?)*$

Fixed in the second edition.

The environment variables used to configure the parse-server application are all missing the PARSE_SERVER_ prefix in sample 14-1.parse.yaml. Updating the Deployment object using the following spec.env section fixes the issue:

---
env:
- name: PARSE_SERVER_DATABASE_URI
value: "mongodb://mongo-0.mongo:27017,\
mongo-1.mongo:27017,mongo-2.mongo\
:27017/dev?replicaSet=rs0"
- name: PARSE_SERVER_APPLICATION_ID
value: my-app-id
- name: PARSE_SERVER_MASTER_KEY
value: my-master-key
---
Reference: https://github.com/parse-community/parse-server/blob/master/README.md#using-environment-variables-to-configure-parse-server

Note from the Author or Editor:
Confirmed, and corrected in the upstream repository.

The text talks about "Once you have this configuration file saved to config.js ..." whereas the example file and command line talk about the file being named "ghost-config.js". So, I think the text should read "Once you have this configuration file saved to ghost-config.js ...".

Note from the Author or Editor:
Confirmed. 'config.js' should be 'ghost-config.js'

The chapter talks about config.js ("...first modify config.js to include:") and also the kubectl command says

kubectl create configmap ghost-config-mysql --from-file config.js

whereas previously, the configuration file was called ghost-config.js, so it would be:

kubectl create configmap ghost-config-mysql --from-file ghost-config.js

Note from the Author or Editor:
Reporter is correct. I fixed this error in the text.

Where you say "Then configure the DHCP server as follows:" you should reference the file you need to edit with the additional information.

i.e. This should read, "Then configure the DHCP server by editing /etc/dhcp/dhcpd.conf as follows:"

Note from the Author or Editor:
Thanks. This has been fixed in the second edition.

After suggesting the edit to /etc/sysctl.conf you need to reboot for the change to take effect.

Note from the Author or Editor:
fixed in source.

In the `kubeadm init` command you use `--api-advertise-addresses` when it should be `--apiserver-advertise-address`

Note from the Author or Editor:
Fixed in second edition.

getting the details of the fluentd daemonSt you say use the command:

kubectl describe demonset fluentd

this returns nothing, however specifying the kube-system namespace does return this daemonset:


kubectl describe daemonset fluentd --namespace=kube-system

Note from the Author or Editor:
Reporter is correct, we should add '--namespace=kube-system' to that command.

There are two commands to curl to get a crt and key in the book it says that these files are created locally however i think you need to write the response to a file

such as::
“curl -O https://storage.googleapis.com/kuar-demo/kuard.crt >> kuard.crt”

Note from the Author or Editor:
Confirmed, the two lines should read:

$ curl -o kuard.crt https://storage.googleapis.com/kuar-demo/kuard.crt
$ curl -o kuard.key https://storage.googleapis.com/kuar-demo/kuard.key

I have fixed this in the Atlas repo.

For the mysql singleton example, both the nfs volume & and nfs volume claim yaml files will not work as written since kubernetes requires you to specify accessModes (spec.accessModes).

Note from the Author or Editor:
Reporter is correct. I have corrected the text in the repository, no further corrections necessary.

In Chapter 13, in the mongo stateful set example...

[$kubectl exec mongo-0 bash ping mongo-1.mongo] will not work with the specified StateFull set definition, the reason being curl does not come pre-installed on the default mongo Docker image. I'm not exactly sure what image the default mongo image is built on top of (e.g. alpine, scratch, wheezy), but in order to get the example to work, I had to change the image spec.template.containers.image value from mongo 3.4.1 to mongo:3.4.9-jessie.

Then running [$kubectl exec -it mongo-0 /bin/bash], I installed curl via [$apt update && apt install iputils-ping].

Then everything else works as listed in the book

Note from the Author or Editor:
The reporter is correct. I have fixed this in the book's original repository. No further action is needed.

Incorrect tabbing of readiness probe portion of the mongo example

Written as: "livenessProbe:\n\texec:\n\t\tcommand:\n\t\t\t- /usr/bin/mongo\n\t\t\t- --eval\n\t\t\t- db.serverStatus()\n\t\tinitialDelaySeconds: 10\n\t\ttimeoutSeconds: 10\n", should be "livenessProbe:\n\texec:\n\t\tcommand:\n\t\t\t- /usr/bin/mongo\n\t\t\t- --eval\n\t\t\t- db.serverStatus()\n\tinitialDelaySeconds: 10\n\ttimeoutSeconds: 10\n"

Note from the Author or Editor:
Reporter is correct.

Tabbing fixed in the origin Atlas repo.