##BEFORE YOU ARRIVE In order to make best use of the lab time, please review the deployment options and ensure either:
- A working KVM environment or
- Access to an OpenStack environment
##Agenda / High Level Overview:
- Deploy Atomic Hosts
- Configure Flannel
- Configure Kubernetes
- Deploy an Application
- Use Super Privileged Containers on the Atomic Hosts
#Deployment There are many ways to deploy an Atomic Host. In this lab, we provide guidance for OpenStack or local KVM.
##Deployment Option 1: Atomic Hosts on OpenStack You may use an OpenStack service, which needs to have a keypair and a security group already configured.
- Navigate to Instances
- Click "Launch Instances" 1 Complete dialog
- Details tab
* Instance name: arbitrary name. Note the UUID of the image will be appended to the instance name. You may want to use your name in the image so you can easily find it.
* Flavor: m1.medium
* Instance count: 3
* Instance Boot Source: Boot from image
- Image name: [atomic_image]
- Access & Security tab * Select your keypair that was uploaded during OpenStack account setup * Security Groups: Default
- Click "Launch"
Three VMs will be created. Once the Power State is Running, you may SSH into the VMs using your matching SSH key.
- Note: Each instance requires a floating IP address in addition to the private OpenStack
172.x.x.x
address. Your OpenStack tenant may automatically assign a floating IP address. If not, you may need to assign it manually. If no floating IP addresses are available, create them.- Navigate to Access & Security
- Click "Floating IPs" tab
- Click "Allocate IPs to project"
- Assign floating IP addresses to each VM instance
- SSH into the VMs with user
cloud-user
and the instance floating IP address. This address will probably be in the10.3.xx.xx
range.
ssh -i <private SSH key> [email protected]
##Deployment Option 2: Atomic Hosts on KVM
- Grab and extract the Atomic and metadata images from our internal repository. Use sudo and appropriate permissions.
wget [metadata ISO image]
cp atomic0-cidata.iso /var/lib/libvirt/images/.
wget [atomic QCOW2 image]
cp rhel-atomic-host-7.qcow2.gz /var/lib/libvirt/images/.; cd /var/lib/libvirt/images
gunzip rhel-atomic-host-7.qcow2.gz
- Make 3 copy-on-write images, using the downloaded image as a "gold" master.
for i in $(seq 3); do qemu-img create -f qcow2 -o backing_file=rhel-atomic-host-7.qcow2 rhel-atomic-host-7-${i}.qcow2 ; done
-
Use the following commands to install the images. Note: You will need to change the bridge (br0) to match your setup, or at least confirm it matches what you have. You may also want to change the os-variant parameter to match your host (e.g. sample values include "rhel7.0", "fedora21", or "centos7.0").
-
For Fedora or CentOS atomic hosts, we need to create cloud init data i.e. atomic0-cidata.iso in below commands. Refer https://www.technovelty.org//linux/running-cloud-images-locally.html
virt-install --import --name atomic-ga-1 --os-variant=rhel7.0 --ram 1024 --vcpus 2 --disk path=/var/lib/libvirt/images/rhel-atomic-host-7-1.qcow2,format=qcow2,bus=virtio --disk path=/var/lib/libvirt/images/atomic0-cidata.iso,device=cdrom,readonly=on --network bridge=br0
virt-install --import --name atomic-ga-2 --os-variant=rhel7.0 --ram 1024 --vcpus 2 --disk path=/var/lib/libvirt/images/rhel-atomic-host-7-2.qcow2,format=qcow2,bus=virtio --disk path=/var/lib/libvirt/images/atomic0-cidata.iso,device=cdrom,readonly=on --network bridge=br0
virt-install --import --name atomic-ga-3 --os-variant=rhel7.0 --ram 1024 --vcpus 2 --disk path=/var/lib/libvirt/images/rhel-atomic-host-7-3.qcow2,format=qcow2,bus=virtio --disk path=/var/lib/libvirt/images/atomic0-cidata.iso,device=cdrom,readonly=on --network bridge=br0
##Update VMs
NOTE: We will be working on all three (3) VMs. You will probably want to have three terminal windows or tabs open.
-
Confirm you can log in to the hosts:
Username: cloud-user Password: atomic (KVM only)
-
Enter sudo shell:
sudo -i
- Update all of the Atomic Hosts. The following commands will subscribe you to receive updates and allow you to upgrade your Atomic Host.
NOTE: Depending on the version of Atomic that you initially installed, some of the sample output below may differ from what you see.
RHEL Atomic Host
# atomic host status
TIMESTAMP (UTC) VERSION ID OSNAME REFSPEC
* 2015-02-17 22:30:38 7.1.244 27baa6dee2 rhel-atomic-host rhel-atomic-host:rhel-atomic-host/7/x86_64/standard
# subscription-manager register --serverurl=[stage] --baseurl=[stage] --username=[account_user] --password=[account_pass] --auto-attach
NOTE: The below output is an example. That is what a customer will see once there is a tree update. What you will see in the lab is that there is "No upgrade available", this is expected.
Fedora Atomic Host
# atomic status
TIMESTAMP (UTC) ID OSNAME REFSPEC
* 2014-12-03 01:30:09 ba7ee9475c fedora-atomic fedora-atomic:fedora-atomic/f21/x86_64/docker-host
NOTE: The below output is an example. That is what a customer will see once there is a tree update. What you will see in the lab is that there is "No upgrade Available", this is expected.
# atomic host upgrade
Updating from: rhel-atomic-host-ostree:rhel-atomic-host/7/x86_64/standard
53 metadata, 321 content objects fetched; 81938 KiB transferred in 71 seconds
Copying /etc changes: 26 modified, 4 removed, 57 added
Transaction complete; bootconfig swap: yes deployment count change: 1
Changed:
libgudev1-208-99.atomic.0.el7.x86_64
libsmbclient-4.1.12-21.el7_1.x86_64
libwbclient-4.1.12-21.el7_1.x86_64
python-six-1.3.0-4.el7.noarch
redhat-release-atomic-host-7.1-20150219.0.atomic.el7.1.x86_64
samba-common-4.1.12-21.el7_1.x86_64
samba-libs-4.1.12-21.el7_1.x86_64
shadow-utils-2:4.1.5.1-18.el7.x86_64
subscription-manager-1.13.22-1.el7.x86_64
subscription-manager-plugin-container-1.13.22-1.el7.x86_64
subscription-manager-plugin-ostree-1.13.22-1.el7.x86_64
systemd-208-99.atomic.0.el7.x86_64
systemd-libs-208-99.atomic.0.el7.x86_64
systemd-sysv-208-99.atomic.0.el7.x86_64
Upgrade prepared for next boot; run "systemctl reboot" to start a reboot
- Check the atomic tree version. The asterisk indicates the currently running tree. The tree displayed first in the list is the version that will be booted into. In the output below, if the system is rebooted, it will boot into the new 7.1.0 tree.
# atomic host status
TIMESTAMP (UTC) VERSION ID OSNAME REFSPEC
2015-02-19 20:26:26 7.1.0 5799825b36 rhel-atomic-host rhel-atomic-host-ostree:rhel-atomic-host/7/x86_64/standard
* 2015-02-17 22:30:38 7.1.244 27baa6dee2 rhel-atomic-host rhel-atomic-host-ostree:rhel-atomic-host/7/x86_64/standard
- Reboot the VMs to switch to the updated tree.
# systemctl reboot
- After the VMs have rebooted, SSH into each and enter the sudo shell:
# sudo -i
- Check your version with the atomic command. The
*
pointer should now be on the new tree.
# atomic host status
TIMESTAMP (UTC) VERSION ID OSNAME REFSPEC
* 2015-02-19 20:26:26 7.1.0 5799825b36 rhel-atomic-host rhel-atomic-host-ostree:rhel-atomic-host/7/x86_64/standard
2015-02-17 22:30:38 7.1.244 27baa6dee2 rhel-atomic-host rhel-atomic-host-ostree:rhel-atomic-host/7/x86_64/standard
Integrating a private registry is an important use case for customers. For this lab, we add a private registry to pull and search images.
- Edit the
/etc/sysconfig/docker
file and restart docker. You will need the following line in the file.
ADD_REGISTRY='--add-registry [PRIVATE_REGISTRY]'
NOTE: If the private registry is not configured with a CA-signed SSL certificate docker pull ...
will fail with a message about an insecure registry. In that case, add the following line to /etc/sysconfig/docker
:
INSECURE_REGISTRY='--insecure-registry [PRIVATE_REGISTRY]'
/etc/sysconfig/docker
includes example ADD_REGISTRY and INSECURE_REGISTRY lines. Uncomment them and append the [PRIVATE_REGISTRY] FQDN. For example:
ADD_REGISTRY='--add-registry my.private.registry.fqdn'
INSECURE_REGISTRY='--insecure-registry my.private.registry.fqdn'
- Restart docker
systemctl restart docker
What can you do? What can't you do? You may see a lot of "Command not found" messages... We'll explain how to get around that with the rhel-tools container in a later lab. Type the following commands.
man tcpdump
git
tcpdump
sosreport
Why wouldn't we include these commands in the Atomic image?
Let's try:
atomic install rhel7/rhel-tools
This will install the rhel-tools container, which can be used as the administrator's shell.
Now let's try:
atomic run rhel7/rhel-tools man tcpdump
atomic run rhel7/rhel-tools tcpdump
You can also go into the rhel-tools container and explore its contents.
atomic run rhel7/rhel-tools /bin/sh
You might even want to create a shell script like the following on the Atomic Host as a helper script:
vi /usr/local/sbin/man
#!/bin/sh
atomic run rhel7/rhel-tools man $@
chmod +x /usr/local/sbin/man
This script makes using man pages transparent to the user (even though man pages are not installed on the Atomic Host, only in the rhel-tools container). It could also be done with a bash alias.
man tcpdump
rhel-tools is a Super Privileged Container, which will be covered in the next presentation and lab.
This concludes the deploying Atomic lab.