EAI-6030: Add AIM_HARDWARE_FAMILY flag for model source selection#256
Open
pre wants to merge 3 commits into
Open
EAI-6030: Add AIM_HARDWARE_FAMILY flag for model source selection#256pre wants to merge 3 commits into
pre wants to merge 3 commits into
Conversation
Add the AIM_HARDWARE_FAMILY install flag (comma-separated: cpu,epyc,instinct,radeon; empty = legacy install-all). A new hardwareFamilyList schema type validates the value via the existing pattern/examples harness. cluster-bloom splits the flag into a YAML list and injects it into cluster-forge's apps.aim-cluster-model-source.valuesObject.hardwareFamilies: via --set-json on the small-cluster ArgoCD render (argocd_deploy.yaml) and via yq into the gitea-persisted values for medium/large (bootstrap_gitea.yaml). Passing a structured list avoids Helm's comma-as-list-separator pitfall entirely, no escaping needed. Bump the LoadSchema field-count assertion (was already stale) to 38. Part of EAI-6030.
6 tasks
The gitea-persistence yq task interpolated the JSON array inside a double-quoted yq expression, so the shell stripped the inner quotes and yq received a bare [epyc, instinct] it could not lex. Single-quote the expression so the JSON string quotes survive to yq. Part of EAI-6030.
The previous yq edit to complete_values.yaml was dropped on medium/large, the gitea-init-job rebuilds cluster-values from a template and only carries over enabledApps plus a fixed global block, so apps.*.valuesObject overrides never reached ArgoCD and the legacy branch rendered. Pass AIM_HARDWARE_FAMILY to the init job via a values file (companion cluster-forge change writes it into cluster-values), matching the disabledApps/airmImageRepository pattern. Small clusters are unaffected (they use --set-json on the direct root render). Part of EAI-6030.
pre
force-pushed
the
EAI-6030-aim-gpu-family
branch
from
342f410 to
1aeab75
Compare
This was referenced Jun 11, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Related
Summary
Adds the
AIM_HARDWARE_FAMILYinstall flag that selects which AIM model sources cluster-forge installs, by hardware family.cpu,epyc,instinct,radeon. Empty (default) = install the full legacy catalog (no behavior change).hardwareFamilyListschema type validates the value via the existing pattern/examples harness (no new Go validator).apps.aim-cluster-model-source.valuesObject.hardwareFamilies:--set-jsonon the ArgoCD render (argocd_deploy.yaml)yqinto the gitea-persisted values (bootstrap_gitea.yaml)Pairs with cluster-forge PR #741 (the chart side). Part of EAI-6030.
Notes
create_cluster_forge_app.yaml, but that task renders only the parent app-of-apps (--show-only templates/cluster-forge.yaml) and cannot carry child-app values. The active small-path applier isargocd_deploy.yaml, so injection lives there instead.LoadSchemafield-count assertion to 38 (it was already stale before this change).Test plan
go build ./...go vet ./pkg/config/...go test ./pkg/config/...(newTestHardwareFamilyListPattern+ validator auto-coverage of valid/invalid examples)[]for empty, correct JSON arrays otherwise (incl. whitespace trim)yqlist write survives the size-merge step