https://github.com/informaticsmatters/galaxy-patch

An Ansible playbook to patch our Galaxy cluster
https://github.com/informaticsmatters/galaxy-patch

Last synced: 9 months ago
JSON representation

An Ansible playbook to patch our Galaxy cluster

• Host: GitHub
• URL: https://github.com/informaticsmatters/galaxy-patch
• Owner: InformaticsMatters
• Created: 2019-10-03T13:50:56.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2021-11-02T12:16:03.000Z (over 4 years ago)
• Last Synced: 2025-01-25T18:43:21.046Z (over 1 year ago)
• Language: Python
• Size: 82 KB
• Stars: 0
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Galaxy Patch (OpenStack)

Ansible playbook and roles to patch an OpenStack Galaxy deployment

This code (two Ansible roles and a `site.yaml`) are used to *increase* the size

of an existing Galaxy/Condor cluster. It simply creates `exec` instances

and then *patches* them by: -

1.  Templating `auto.data`. **Importantly** The NFS exports must already be

    available - this play does not setup the NFS server's devices, mounts or

    exports

2.  Templating `condor_config.local`

3.  Templating `data.autofs`

4.  Restarting the `nfs`, `autofs` and `condor` services

### Significant variables

-   `exec_base_name`

-   `exec_type`

-   `exec_count` - the number of new exec instances to create

-   `exec_offset` - the numerical offset of the first exec instance.

    If you're adding to a cluster and you have instances `0` to `10`

    this would normally be `11`

-   `create_exec_instances` - if `no` new instances are not created

    but the *patching* is still applied to all exec instances in the cluster

## Usage

1.  Clone onto a suitable *bastion* in the cluster

2.  Inspect the variables (in `group_vars/all/main.yaml`) and create

    a local `parameters.yaml` file for any values you need to adjust

3.  Inspect the templates to ensure they reflect your needs

    (specifically the mounts in `auto.data.j2`)

Normally run from the DLS bastion as `mauro`: -

    $ sudo -iu mauro

    $ cd galaxy-patch

    $ git pull

Set your keystone environment variables (here we're using a script that 

prompts you for your username and password), check the parameters,

and run the playbook with: -

 

    $ source ~/keystone_v3_for_anyone.sh

    [...]

    [check parameters.yaml] 

    $ ansible-playbook site.yaml -e @parameters.yaml

>   You may need to run the playbook a number of times - especially if your

    instances have different base names. At the time of writing we have

    used `pulsar-exec-node`, `pulsar-exec-node-b`, `pulsar-exec-node-gpu`,

    `pulsar-exec-node-gpu-b` and `pulsar-submit-machine` as a base-names

    for our instances.

>   The machines that match your chosen `exec_base_name`

    are expected to be named `-[0-9]*`. If the machine names

    do not match, no machines will be placed in the inventory and nothing

    will be run.

For all our current machines, using the command-line for non-GPU nodes: -

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-a -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-b -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-c -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-d -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-e -e exec_is_gpu=no

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-submit-machine -e exec_is_gpu=no

and for GPU nodes: -

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-cuda -e exec_is_gpu=yes

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-cuda-a -e exec_is_gpu=yes

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-cuda-b -e exec_is_gpu=yes

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-cuda-x4 -e exec_is_gpu=yes

    $ ansible-playbook site.yaml -e exec_base_name=pulsar-exec-node-cuda-x2 -e exec_is_gpu=yes

---