azure-gate: deploy via az rest (nixpkgs az deployment module crashes on NixOS) #154
No reviewers
Labels
No labels
bug
discussion
duplicate
enhancement
goal
help wanted
horizon:backlog
horizon:governance
horizon:mvp
invalid
operator-decision
question
roadmap
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
open-platform/mitosis!154
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "fix/azure-gate-az-rest"
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?
The blocker after the bicep fix (#151). Diagnosed from Forgejo's action-log storage and reproduced on the runner: even with
nixpkgs#bicepon PATH anduse_binary_from_path=true,az deployment group createcrashes on the NixOS runner:The nixpkgs azure-cli's
az deploymentcommand module is broken (same class as theaz aks nodepool addSDK crash documented inaks-cost-experiment.md). Confirmed on vxrail that BOTH the.bicepand a pre-compiled.jsoncrash identically and early → it's the command module, not bicep.az restagainst the ARM deployments API works (validated on the runner:provisioningState: Succeeded).Fix (
deploy/azure/gate.sh+ workflow):deploy()helper: compile the bicep to ARM JSON with nix bicep, submit viaaz restPUT, pollprovisioningState;dep_output()reads outputs viaaz restGET. Both vms + aks deploy sites use it.bicepon PATH) keep usingaz deployment group createunchanged.nixpkgs#python3(deploy() builds the ARM payload with it).Note: this supersedes Ben's #153 (which is the same bicep fix already merged as #151) — #153 can be closed. This one goes a layer deeper to the actual deploy crash.
🤖 Generated with Claude Code