Skip to content

Latest commit

 

History

History
42 lines (31 loc) · 2.6 KB

README.md

File metadata and controls

42 lines (31 loc) · 2.6 KB
Raw

AIS K8s Playbooks

A small set of convenient playbooks to assist in preparing K8s worker nodes to host an AIStore deployment. None of these are required. We use all of them in our reference environment, but you're free to make the filesystem as you wish, tune nodes as you wish, etc., in which case either ignore these or use them as a reference.

These playbooks have been tested only on Ubuntu hosts.

For playbooks related to cloud configuration, see cloud

Each playbook is documented separately. See the links in the first column below.

Playbook(s) Useful when
ais_enable_multiqueue Enabling MQ IO schedulers in Ubuntu releases for which MQ is not the default
ais_host_config_common Tuning worker nodes
ais_datafs_mkfs Creating or recreating filesystems for AIStore
config_kubelet Enabling K8s settings that must be set at a kubelet service level, e.g. unsafe sysctls
ais_gpuhost_config Configuring GPU compute nodes in the same cluster - install NVIDIA Docker 2, NVIDIA container runtime, etc.
ais_gpu_host_config (EXPERIMENTAL) Additional setup to use Nvidia GPUs on hosts. This playbook is experimental, so check the roles and use at your own risk.

The ais_host_config_common playbook includes a tagging scheme to allow more granular selection of tasks and to skip tasks that are likely site-specific.

The vars directory includes variable definitions that control the playbooks, split into multiple files with comments explaining which playbooks they control (and which tags will use them).

Playbook Order

We run the host config playbooks in the following order wrt other steps:

  1. Cluster hosts must be linux, preferrably Ubuntu >18.04 with ssh access from the ansible host.
  2. If necessary, enable MQ IO scheduler with ais_enable_multiqueue and reboot.
  3. Next we run ais_host_config_common on all nodes. Check the tags in the task to see what's avaiable. At the least, run the playbook with the aisrequired tag.
  4. If we're install gpu worker nodes, run ais_gpuhost_config.
  5. Make filesystems with ais_datafs_mkfs.
  6. Establish K8s cluster using kubespray or other methods, e.g. kubeadm.
  7. Allow "unsafe" network sysctls and any other required kubelet settings with config_kubelet.

At this point your cluster should be ready to deploy the AIS resources, operator, and cluster. See the ais-deployment section and guide