PowerEdge R720 QEMU & Kubernetes Setup #23
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "r720-setup"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Part 1
In this video I am attempting to recycle the
main.yaml
Ansible playbook from this repository to be used for the PowerEdge R720 server. The goal here is to deploy five additional virtual machines to the existing Kubernetes cluster to alleviate recent compute/disk pressure. To my surprise, the original code for deploying virtual machines was not copied to DevOps/ansible-role-libvirt-guest and is now hidden in the commit history. I was able to find a non-functional version in another file, and eventually discovered the original in commit v1.0.0, but this problem prevented the full completion of the PR in one sitting; even so, only a small amount of work remains to fully debug the new playbook. Once completed, I will have several follow-up tasks to complete, which have been documented in the following issues:Additionally, the original issue being addressed here, DevOps/ansible-role-eom#37, will not be completed until the new NFS client is utilized, as I believe the service crashes are due to disk I/O limitations rather than insufficient compute power.
PowerEdge R720 QEMU & Kubernetes Setupto WIP: PowerEdge R720 QEMU & Kubernetes SetupPart 2
Following up from yesterday's live stream, I was able to get further into the deployment by referencing the code from commit 6e7ee42c1e; unfortunately, the playbook subsequently failed due to networking issues. I believe this was caused by the assignment of a new IP address to the virtual bridge created for the
wan
libvirt network. Once this resource was created, the host and its guests needed to be accessed via a new address. Having rebooted the R720, I believe this is now fixed and will be able to continue in the next stream.WIP: PowerEdge R720 QEMU & Kubernetes Setupto PowerEdge R720 QEMU & Kubernetes SetupPart 3
Success! After rebooting the hypervisor and rerunning the playbook, I was able to confirm the addition of 5 nodes to the Kubernetes cluster. As mentioned in the video, I do not expect this to solve the issue presented in DevOps/ansible-role-eom#37, but I will address that in another PR after ensuring my assumptions about needing to use the second NFS client are confirmed to be correct. Please note that while this PR is merged and, therefore, technically closed, users are invited to continue posting questions or suggestions here (as well as on any other issue/pull request).