Refactor: Fix broken persys-scheduler / Implement compute agent / reconciler + scheduler

Author: miladhzzzz

persys-scheduler/Dockerfile

@@ -0,0 +1,30 @@
+# Stage 1: Build the application
+FROM golang:1.24-alpine AS builder
+
+# Set the working directory
+WORKDIR /app
+
+# Copy go.mod and go.sum to download dependencies
+COPY go.mod go.sum ./
+RUN go mod download
+
+# Copy the rest of the application code
+COPY . .
+
+# Build the application
+RUN CGO_ENABLED=0 GOOS=linux go build -o persys-scheduler ./cmd/scheduler
+
+# Stage 2: Create the final image
+FROM alpine:latest
+
+# Set the working directory
+WORKDIR /root/
+
+# Copy the binary from the builder stage
+COPY --from=builder /app/persys-scheduler .
+
+# Expose the port the scheduler listens on
+EXPOSE 8084
+
+# Command to run the scheduler
+CMD ["./persys-scheduler"]

persys-scheduler/Makefile

@@ -0,0 +1,102 @@
+# Makefile for Go project
+
+# Variables
+BINARY_NAME = persys-scheduler
+BINARY_DIR = bin
+BINARY_PATH = $(BINARY_DIR)/$(BINARY_NAME)
+IMAGE_NAME=persys-scheduler
+IMAGE_TAG=latest
+GO = go
+GOFLAGS = -v
+PORT = 8084
+
+# Default target
+.PHONY: all
+all: build
+
+# Ensure bin directory exists
+$(BINARY_DIR):
+	mkdir -p $(BINARY_DIR)
+
+# Build the binary into bin directory
+.PHONY: build
+build: $(BINARY_DIR)
+	$(GO) build $(GOFLAGS) -o $(BINARY_PATH) ./cmd/scheduler
+
+# Build the Docker image
+.PHONY: docker-build
+docker-build:
+	$(DOCKER) build -t $(IMAGE_NAME):$(IMAGE_TAG) .
+	@echo "Built Docker image $(IMAGE_NAME):$(IMAGE_TAG)"
+
+# Run the PersysAgent container
+.PHONY: docker-run
+docker-run:
+	$(DOCKER) compose up -d
+	@echo "Started $(BINARY_NAME) container, API on prow:8084"
+
+# Run the application from bin directory
+.PHONY: run
+run: build
+	./$(BINARY_PATH)
+
+# Test the code (if you add tests later)
+.PHONY: test
+test:
+	$(GO) test $(GOFLAGS) ./...
+
+# Clean up generated files
+.PHONY: clean
+clean:
+	$(GO) clean
+	rm -rf $(BINARY_DIR)
+
+# Format the code
+.PHONY: fmt
+fmt:
+	$(GO) fmt ./...
+
+# Vet the code for potential issues
+.PHONY: vet
+vet:
+	$(GO) vet ./...
+
+# Update dependencies
+.PHONY: deps
+deps:
+	$(GO) mod tidy
+	$(GO) mod download
+
+# Build and run with a single command from bin directory
+.PHONY: dev
+dev: build
+	./$(BINARY_PATH)
+
+# Check for linting issues (requires golangci-lint)
+.PHONY: lint
+lint:
+	golangci-lint run
+
+# Install the binary to $GOPATH/bin
+.PHONY: install
+install:
+	$(GO) install $(GOFLAGS)
+
+# Help command to display available targets
+.PHONY: help
+help:
+	@echo "Available targets:"
+	@echo "  all      	   - Build the project into bin/ (default)"
+	@echo "  build    	   - Build the binary into bin/"
+	@echo "  docker-build  - Build Docker image"
+	@echo "  docker-run    - Run PersysAgent container"
+	@echo "  run           - Build and run the application from bin/"
+	@echo "  test          - Run tests"
+	@echo "  clean         - Remove generated files and bin/ directory"
+	@echo "  fmt           - Format the code"
+	@echo "  vet           - Vet the code"
+	@echo "  deps          - Update and download dependencies"
+	@echo "  dev           - Build and run from bin/ for development"
+	@echo "  lint          - Run linter (requires golangci-lint)"
+	@echo "  install       - Install the binary to $$GOPATH/bin"
+	@echo "  help          - Show this help message"

persys-scheduler/README.md

@@ -0,0 +1,77 @@
+# Persys Scheduler
+
+`persys-scheduler` is the scheduler/control-plane service for Persys workloads.
+
+## Responsibilities
+
+- Accept node registration + heartbeat over gRPC.
+- Persist node/workload state in etcd.
+- Select nodes for workloads based on status, labels, capabilities, and available resources.
+- Reconcile desired state (`Running`/`Stopped`/`Deleted`) with actual state.
+- Track retries, assignments, reconciliation records, and events.
+
+## API
+
+Proto: `api/proto/control.proto`
+Service: `persys.control.v1.AgentControl`
+
+Implemented RPCs include:
+
+- `RegisterNode`
+- `Heartbeat`
+- `ApplyWorkload`
+- `DeleteWorkload`
+- `RetryWorkload`
+- `ListNodes`
+- `GetNode`
+- `ListWorkloads`
+- `GetWorkload`
+- `GetClusterSummary`
+
+## Default ports
+
+- Metrics + health HTTP: `:8084`
+- gRPC control plane: `:8085`
+
+Health endpoints:
+
+- `GET /health`
+- `GET /metrics`
+
+## Local run
+
+Start etcd first, then run scheduler.
+
+```bash
+./hack/start-etcd-docker.sh
+cd persys-scheduler
+go run ./cmd/scheduler -insecure
+```
+
+`-insecure` disables mTLS for local testing.
+
+## Key environment variables
+
+See `sample.env`.
+
+Most used:
+
+- `ETCD_ENDPOINTS` (default `localhost:2379`)
+- `DOMAIN` (default `persys.local`)
+- `GRPC_PORT` (default `8085`)
+- `METRICS_PORT` (default `8084`)
+
+For mTLS mode:
+
+- `CA_FILE`
+- `CFSSL_API_URL`
+- `CERT_COMMON_NAME`
+- `CERT_ORGANIZATION`
+
+## Build and test
+
+```bash
+cd persys-scheduler
+go build ./cmd/scheduler
+go test ./...
+```