Note: If any error occurs, there are many ways to debug. It's possible to go through the logs directory for ERRORS that may come up.
/var/logs/<service>
Most of the items are based on the installation instruction in the OpenStack Ocata Page linked below:
OpenStack Official Documentation ( Ocata )
Install vbox additions then mount
mkdir ~/GuestShared
sudo mount -t vboxsf -o uid=1000,gid=1000,rw GuestShared ~/GuestShared
/etc/hostnameChange the file to the hostname of the node/etc/hostsChange hosts to point into particular nodes in
Change the interfaces in network/interfaces within each folder. Refer to the mapping below:
controller 10.x.x.10
compute 10.x.x.20
store 10.x.x.30
network 10.x.x.40
3 adapters will be used in the system configuration: Data, Management and External NAT network to access the internet as setup below:
Management 10.10.10.x
Data 10.20.20.x
NatNetwork 10.0.2.x
NTP is used to synchronize all time related within the nodes. We are setting Chrony on controller and syncing the rest to the controller node.
Controller + Other nodes
Refer to addSources file
Refer to MySQL Folder
OpenStack uses a messaging queue for services to coordinate with each other. There are many choices available but we will be using RabbitMQ
Check folder RabbitMQ
Memcached is used to cache tokens whenever Keystone tries to authenticate a user.
Configure memcached on the controller. Find the flag -l and add the controller on management network.
-l 10.10.10.10
Restart the memcached
sudo service memcached restart
Check folder Memcached
Check folder Keystone
After running the 01 install script, do the following:
Add/change in file /etc/keystone/keystone.conf
[database]
# ...
connection = mysql+pymysql://keystone:letmein@controller/keystone
[token]
# ...
provider = fernet
Run 02 install script
Set the servername in /etc/apache2/apache2.conf on top of the page like so:
ServerName controller
Check Keystone/testing folder.
Execute the command
. admin-openrc
openstack token issue
Try the demo as well
Sets up the system as a repository of images that is stored as a file in the default directory /var/lib/glance/images/
Check Glance folder
Run installGlance
Edit the file /etc/glance/glance-api.conf and add the following
[database]
# ...
connection = mysql+pymysql://glance:letmein@controller/glance
[keystone_authtoken]
# ...
auth_uri = http://controller:5000
auth_url = http://controller:35357
memcached_servers = controller:11211
auth_type = password
project_domain_name = default
user_domain_name = default
project_name = service
username = glance
password = letmein
[paste_deploy]
# ...
flavor = keystone
[glance_store]
# ...
stores = file,http
default_store = file
filesystem_store_datadir = /var/lib/glance/images/
There's additional things to do. Refer to the official documentation
After that, do postInstallGlance
If error comes up regarding locale, run
export LC_ALL=en_US.UTF-8
Check install instructions inside ComputeConf in controller and Compute in compute node
Again, after running installComputeConf, there will still be addition configurations for it. Refer to the official documentations
Check compute folder inside compute node.
After installing using the installNova, continue configuration from the official documentations
Networking is one of the challenging tasks for there are 2 choices of which the installation can continue. The networking module can be set aside in its own node in the system. However, the decided option to do is to put them inside the controller node as the project itself is not complicated enough to balance the load of the controller.
/// EDIT HERE
This is the chosen option as it gives enough flexibility on the tenant users to create their own networking configuration that suits their needs. Compared to the first option, this is a more powerful way of managing networking as it extends beyond what the first option offers, which is simply adding the instance in the network and that's about it. Option 2 adds a layer of functionality that will almost always be needed.
Refer to the Networking folder (controller node)
After the installNetworking script is done, refer to the official installation on the configurations.
Since Option 2 is used, option 2 configuration should be done depending on the instructions writted on the official documentation
Refer to the Networking folder.
Since Option 2 is used, option 2 configuration should be done depending on the instructions writted on the official documentation
Setup provider network first, then selfservice network. After, create a router using demo account
Run ip netns with adminrc to show possible namespaces
Trying to ping the router needs to have the namespace of the ip to be the dhcp of the router. And so, a command like below is to be ran to execute the ping command properly:
sudo ip netns exec qdhcp-22d0f59b-71ec-47c3-b8c2-f3b883eeb8c4 ping 203.0.113.109
The address 203.0.113.109 of the router is listed from the command
neutron router-port-list router
m1.nano flavor
virtual cpu = 1 ram = 64 disks = 1 name: m1.nano
Launching the instance will need the selfservice net id which can be taken from the command
openstack network list
When restarting and trying to access the instances, make sure that the network services and servers are active. To check,
openstack server list
It shows the servers that outputs like so:
| ID | Name | Status | Networks | Image Name |
|---|---|---|---|---|
| 1c376f83-bd1a-410d-9978-0f5d2c676d7b | provider-instance | BUILD | cirros | |
| ea9e26ff-0bb1-43bd-94f8-0eeac609a6f2 | selfservice-instance | ACTIVE | selfservice=172.16.1.11 | cirros |
Make sure that it's ACTIVE before going beyond. The names above are the name
of the instances. Incase the STATUS is SHUTOFF, start them by doing the
commands below:
nova start <name>
## i.e.
## nova start selfservice-instanceList the servers again and check if it's active. It may take awhile.
Check nova services. Restart the services if any error occurs when accessing
the vnc