adoptium / adoptium/infrastructure
Discussion: Decommission bare metal macstadium machines - move funding into Orka x64 cluster
- Dominant language
- Python
- Stars
- 96
- Forks
- 106
- Avg merge
- 1d 23h
- Merged PRs (30d)
- 13
Description
https://github.com/adoptium/infrastructure/issues/4267 is where I am upgrading test-macstadium-macos1014-x64-1 to a MacOS12 machine (for no extra cost in macstadium). However since MacOS12 is EOL, I am hitting various playbook errors, mostly with installing packages via brew that arent there.
So instead of upgrading the 1014 machine, we can decommission it and move the funding into the x64 Okra cluster. At the moment we have more arm64 orka nodes than x64
```
NAME AVAILABLE-CPU AVAILABLE-MEMORY STATUS
arm-mini-003 4 9.60G READY
arm-mini-004 8 16.00G READY
arm-mini-005 8 16.00G READY
arm-mini-006 8 16.00G READY
arm-mini-007 8 16.00G READY
arm-mini-008 8 16.00G READY
x86-mini-1 0 6.26G READY
x86-mini-2 0 6.26G READY
```
These nodes are host nodes on which our dynamic orka nodes run. At the moment we are bottlenecked in our x64 capacity so we would benefit from an increase in x64 worker nodes
Contributor guide
Research direction
Start with issue #4267 and review the playbook errors encountered while upgrading test-macstadium-macos1014-x64-1. Compare the current x86 host capacity with the arm64 nodes described here. Done means deciding whether to decommission the bare-metal machine and documenting or implementing the move to additional x64 Orka capacity.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- ansible, macos
- Domain
- devops, infrastructure
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100