vfarcic / cloud-provisioning Goto Github PK

Hi, I am reading your book that devops-toolkit 2.1.

At "Collecting Metrics and Monitoring The Cluster"
I am trying your example but Prometheus Target - cadvisor and node Enpoints are wrong. I think beacuse of dns.

I am using latest version of cloud-provisioning repo.

I attached the screenshots.

Is there anything missed?

@vfarcic As you already mention in your book The Devops 2.1 Toolkit, chapter "Setting Up Continuous Delivery Clusters", there is an issue with the volume mounts in current versions of docker-machine and VirtualBox on Windows.
Here's what I found out on this subject using Docker Toolbox 1.13.1 (i.e. Docker Machine 0.9.0 & VirtualBox 5.1.14) on Windows 7:

Your code uses the current directory mounted as a VirtualBox shared folder in several VMs in order to share registry data between the testing and production swarms.
Sadly, mounting custom directories as shared folders on Windows is broken in docker-machine 0.9.0.
Only the default mount of the user's home directory works.

This leads to two problems:

• The VMs do not contain the path $PWD. docker service create fails because of an invalid mount option.
• No data is shared between the swarms.

To verify the shared directory issue, create two machines (xx is a substitute for your user name):

docker-machine create --driver virtualbox default-shared-folder
docker-machine create --driver virtualbox --virtualbox-share-folder "/c/Users/xx:/XXX" custom-shared-folder

In the Oracle VM VirtualBox Manager, check the VM details. default-shared-folder has one shared folder, custom-shared-folder has none.

The reason for the failed creation of shared folders is that VIRTUALBOX_SHARE_FOLDER and --virtualbox-share-folder both have a source:destination syntax, where source on Windows machines must be given as c:/.... The extra colon causes parsing of the argument to fail, see docker/machine#3975. This is fixed in docker/machine#3975. Using a custom build of docker-machine, I was able to create a custom mount of the current directory with

docker-machine create --driver virtualbox --virtualbox-share-folder "$(cygpath -m $PWD):$PWD" custom-shared-folder

Because $pwd returns the path in /c/... format, some kind of conversion is needed. This example will not work on Macs. We need OS specific treatment.

I see the following solutions:

• create separate scripts for Windows and Mac, the Windows scripts using conversion with cygpath (available in Docker Toolbox) or variable substitutions.
• enhance the existing scripts with OS specific conversion.
• create an issue in Docker-Machine that on Windows, paths like /c/... should be converted to c:/....
• Remove VIRTUALBOX_SHARE_FOLDER="$PWD:$PWD" from the scripts. All users will have to work in a subdir of their homedir (like in vfarcic/go-demo). This is the only solution that works with the current Docker Machine.

Hope this information is helpful for you.

I have run the Packer Build command against the various Packer json files in the "terraform/aws" and the "terrform/aws-full" directories and I encountered the same error with each one... resulting in build termination and no AMI image existing for any of them:

Error:
"amazon-ebs: E: can not open /var/lib/apt/lists/archive.ubuntu.com_ubuntu_dists_xenial_InRelease - fopen (2: No such file or directory)"

Recommend:
Upon adding the "sudo apt-get clean" command to the "provisioners : inline" section for each of the Packer json files I was able to get them to build correctly.

Example:
"provisioners": [{ "type": "shell", "inline": [ "sudo apt-get clean", "sudo apt-get update", "sudo apt-get install -y apt-transport-https ca-certificates nfs-common", "sudo apt-key adv --keyserver hkp://ha.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D", "echo 'deb https://apt.dockerproject.org/repo ubuntu-xenial main' | sudo tee /etc/apt/sources.list.d/docker.list", "sudo apt-get update", "sudo apt-get install -y docker-engine", "sudo usermod -aG docker ubuntu", "sudo curl -L \"https://github.com/docker/compose/releases/download/1.8.1/docker-compose-$(uname -s)-$(uname -m)\" -o /usr/local/bin/docker-compose", "sudo chmod +x /usr/local/bin/docker-compose" ] }]

I'm following along with the leanpub DevOps2.1 Toolkit. So far so good, but now the jenkins-agent service creating is proving difficult. With the swarm running:

export USER=admin
export PASSWORD=admin

docker service create --name jenkins-agent \
-e COMMAND_OPTIONS="-master \
http://$(docker-machine ip swarm-1):8082/jenkins \
-username $USER -password $PASSWORD \
-labels 'docker' -executors 5" \
--mode global \
--constraint 'node.labels.env == jenkins-agent' \
--mount "type=bind,source=/var/run/docker.sock,target=/var/run/docker.sock" \
--mount "type=bind,source=$HOME/.docker/machine/machines,target=/machines" \
--mount "type=bind,source=/workspace,target=/workspace" \
vfarcic/jenkins-swarm-agent

If the line --mount "type=bind,source=$HOME/.docker/machine/machines,target=/machines" \ exist, it 'll show ERROR "invalid mount config for type…"

[weblogic@robertleepc cloud-provisioning]$ docker service  ps jenkins-agent 
ID            NAME                                         IMAGE                               NODE          DESIRED STATE  CURRENT STATE           ERROR                             PORTS
s1za0gvjjrri  jenkins-agent.meb853xkpo07kzz618yct2vua      vfarcic/jenkins-swarm-agent:latest  swarm-test-1  Ready          Rejected 3 seconds ago  "invalid mount config for type…"  
6ik84sdnmml6   \_ jenkins-agent.meb853xkpo07kzz618yct2vua  vfarcic/jenkins-swarm-agent:latest  swarm-test-1  Shutdown       Rejected 8 seconds ago  "invalid mount config for type…"  
sw49sebngqv9   \_ jenkins-agent.meb853xkpo07kzz618yct2vua  vfarcic/jenkins-swarm-agent:latest  swarm-test-1  Shutdown       Rejected 8 seconds ago  "invalid mount config for type…"  

I need the line --mount "type=bind,source=$HOME/.docker/machine/machines,target=/machines" \ exist, it 'll show ERROR ``` in order to let Jenkins execute the Job "Pipeline go-demo"(scripts/go-demo-pull.groovy)!

withEnv([
        "DOCKER_TLS_VERIFY=1",
        "DOCKER_HOST=tcp://${env.PROD_LIKE_IP}:2376",
        "DOCKER_CERT_PATH=/machines/${env.PROD_LIKE_NAME}"
])

I'm using Centos 7.3 ( 3.10.0-514.10.2.el7.x86_64)
Version info:

[weblogic@robertleepc cloud-provisioning]$ docker version
Client:
 Version:      17.03.1-ce
 API version:  1.27
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Fri Mar 24 00:36:45 2017
 OS/Arch:      linux/amd64

Server:
 Version:      17.03.1-ce
 API version:  1.27 (minimum version 1.12)
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Mon Mar 27 16:58:30 2017
 OS/Arch:      linux/amd64
 Experimental: false

I've tried this on my mac and on my linux box.

In both cases it just hangs at "waiting for an IP" and the docker-machine is never provisioned.

The following fails for the current Digital Ocean terraform source:

$ terraform plan
...
Error running plan: 2 error(s) occurred:

• Resource 'digitalocean_droplet.swarm-manager' not found for variable 'digitalocean_droplet.swarm-manager.2.id'
• Resource 'digitalocean_droplet.swarm-manager' not found for variable 'digitalocean_droplet.swarm-manager.1.id'

Hi Viktor,

As always, Thank you sir for your hard work and great books. I've run into an issue though...

The first instruction in the "Using Packer To Create Amazon Machine Images (AMIs)" section of the book is to change to a specific directory [from root: vfarcic/cloud-provisioning]

cd terraform/aws

However, the 'aws' directory doesn't exist. Is this a typo, or, is a directory missing?

https://www.terraform.io/docs/providers/aws/r/eip.html

It'd be great if we could have elastic IPs created as part of the terraform scripts, that way it wouldn't be a manual step after the cluster's creation.

When deploying swarm using terraform v0.9.1 output variables are not showing but with v0.8.5 everything works as it show do.

After deploying the first manager node with
terraform apply \ -target aws_instance.swarm-manager \ -var swarm_init=true \ -var swarm_managers=1
No output variables are displayed.

I'm following along with the leanpub DevOps2.1 Toolkit. So far so good, but now the Jenkins CI/CD setup is proving difficult. With the swarm running:

Let’s create a Jenkins service.

1 mkdir -p docker/jenkins
2
3 docker service create --name jenkins \
4 -p 8082:8080 \
5 -p 50000:50000 \
6 -e JENKINS_OPTS="--prefix=/jenkins" \
7 --mount "type=bind,source=$PWD/docker/jenkins,target=/var/jenkins_home" \
8 --reserve-memory 300m \
9 jenkins:2.7.4-alpine
10
11 docker service ps jenkins

...
docker service ps jenkins reports an error for the service ("invalid bind mount source, source path not found..."). However, the directory $PWD/docker/jenkins exists on local host.

I find a reference to this issue here: https://github.com/docker/docker/issues/25912. Still I can't resolve it.

Feels like I'm missing something obvious... Any help greatly appreciated. Thanks for sharing your work.