3fb6902fa1
Update worker system for scheduler integration: - Worker server with scheduler registration - Configuration with scheduler endpoint support - Artifact handling with integrity verification - Container executor with supply chain validation - Local executor enhancements - GPU detection improvements (cross-platform) - Error handling with execution context - Factory pattern for executor instantiation - Hash integrity with native library support
106 lines
2.7 KiB
Go
106 lines
2.7 KiB
Go
// Package errors provides structured error types for the worker.
|
|
package errors
|
|
|
|
import (
|
|
"fmt"
|
|
"time"
|
|
)
|
|
|
|
// ExecutionError provides structured error context for task execution failures.
|
|
// It captures the task ID, execution phase, specific operation, root cause,
|
|
// and additional context to make debugging easier.
|
|
type ExecutionError struct {
|
|
Timestamp time.Time
|
|
Cause error
|
|
Context map[string]string
|
|
TaskID string
|
|
Phase string
|
|
Operation string
|
|
}
|
|
|
|
// Error implements the error interface with a formatted message.
|
|
func (e ExecutionError) Error() string {
|
|
return fmt.Sprintf("[%s/%s] task=%s: %v", e.Phase, e.Operation, e.TaskID, e.Cause)
|
|
}
|
|
|
|
// Unwrap returns the underlying error for error chain inspection.
|
|
func (e ExecutionError) Unwrap() error {
|
|
return e.Cause
|
|
}
|
|
|
|
// WithContext adds a key-value pair to the error context.
|
|
func (e ExecutionError) WithContext(key, value string) ExecutionError {
|
|
if e.Context == nil {
|
|
e.Context = make(map[string]string)
|
|
}
|
|
e.Context[key] = value
|
|
return e
|
|
}
|
|
|
|
// ContextString returns a formatted string of all context values.
|
|
func (e ExecutionError) ContextString() string {
|
|
if len(e.Context) == 0 {
|
|
return ""
|
|
}
|
|
result := ""
|
|
for k, v := range e.Context {
|
|
if result != "" {
|
|
result += ", "
|
|
}
|
|
result += fmt.Sprintf("%s=%s", k, v)
|
|
}
|
|
return result
|
|
}
|
|
|
|
// NewExecutionError creates a new ExecutionError with the given parameters.
|
|
func NewExecutionError(taskID, phase, operation string, cause error) ExecutionError {
|
|
return ExecutionError{
|
|
TaskID: taskID,
|
|
Phase: phase,
|
|
Operation: operation,
|
|
Cause: cause,
|
|
Context: make(map[string]string),
|
|
Timestamp: time.Now(),
|
|
}
|
|
}
|
|
|
|
// IsExecutionError checks if an error is an ExecutionError.
|
|
func IsExecutionError(err error) bool {
|
|
_, ok := err.(ExecutionError)
|
|
return ok
|
|
}
|
|
|
|
// Common error operations for the worker lifecycle.
|
|
const (
|
|
// Preparation operations
|
|
OpCreateWorkspace = "create_workspace"
|
|
OpFetchDataset = "fetch_dataset"
|
|
OpMountVolume = "mount_volume"
|
|
OpSetupEnvironment = "setup_environment"
|
|
OpStageSnapshot = "stage_snapshot"
|
|
|
|
// Execution operations
|
|
OpStartContainer = "start_container"
|
|
OpExecuteCommand = "execute_command"
|
|
OpMonitorExecution = "monitor_execution"
|
|
|
|
// Collection operations
|
|
OpCollectResults = "collect_results"
|
|
OpUploadArtifacts = "upload_artifacts"
|
|
OpCleanupWorkspace = "cleanup_workspace"
|
|
|
|
// Validation operations
|
|
OpValidateManifest = "validate_manifest"
|
|
OpCheckResources = "check_resources"
|
|
OpVerifyProvenance = "verify_provenance"
|
|
)
|
|
|
|
// Common phase names (should match TaskState values).
|
|
const (
|
|
PhaseQueued = "queued"
|
|
PhasePreparing = "preparing"
|
|
PhaseRunning = "running"
|
|
PhaseCollecting = "collecting"
|
|
PhaseCompleted = "completed"
|
|
PhaseFailed = "failed"
|
|
)
|