feat(agent): add script data dir for binaries and files (#12205)

The agent is extended with a `--script-data-dir` flag, defaulting to the
OS temp dir. This dir is used for storing `coder-script-data/bin` and
`coder-script/[script uuid]`. The former is a place for all scripts to
place executable binaries that will be available by other scripts, SSH
sessions, etc. The latter is a place for the script to store files.

Since we default to OS temp dir, files are ephemeral by default. In the
future, we may consider adding new env vars or changing the default
storage location. Workspace startup speed could potentially benefit from
scripts being able to skip steps that require downloading software. We
may also extend this with more env variables (e.g. persistent storage in
HOME).

Fixes #11131

agent/agentscripts/agentscripts.go

@@ -43,11 +43,12 @@ var (
 
 // Options are a set of options for the runner.
 type Options struct {
-	LogDir     string
-	Logger     slog.Logger
-	SSHServer  *agentssh.Server
-	Filesystem afero.Fs
-	PatchLogs  func(ctx context.Context, req agentsdk.PatchLogs) error
+	DataDirBase string
+	LogDir      string
+	Logger      slog.Logger
+	SSHServer   *agentssh.Server
+	Filesystem  afero.Fs
+	PatchLogs   func(ctx context.Context, req agentsdk.PatchLogs) error
 }
 
 // New creates a runner for the provided scripts.
@@ -59,6 +60,7 @@ func New(opts Options) *Runner {
 		cronCtxCancel: cronCtxCancel,
 		cron:          cron.New(cron.WithParser(parser)),
 		closed:        make(chan struct{}),
+		dataDir:       filepath.Join(opts.DataDirBase, "coder-script-data"),
 		scriptsExecuted: prometheus.NewCounterVec(prometheus.CounterOpts{
 			Namespace: "agent",
 			Subsystem: "scripts",
@@ -78,6 +80,7 @@ type Runner struct {
 	cron          *cron.Cron
 	initialized   atomic.Bool
 	scripts       []codersdk.WorkspaceAgentScript
+	dataDir       string
 
 	// scriptsExecuted includes all scripts executed by the workspace agent. Agents
 	// execute startup scripts, and scripts on a cron schedule. Both will increment
@@ -85,6 +88,17 @@ type Runner struct {
 	scriptsExecuted *prometheus.CounterVec
 }
 
+// DataDir returns the directory where scripts data is stored.
+func (r *Runner) DataDir() string {
+	return r.dataDir
+}
+
+// ScriptBinDir returns the directory where scripts can store executable
+// binaries.
+func (r *Runner) ScriptBinDir() string {
+	return filepath.Join(r.dataDir, "bin")
+}
+
 func (r *Runner) RegisterMetrics(reg prometheus.Registerer) {
 	if reg == nil {
 		// If no registry, do nothing.
@@ -104,6 +118,11 @@ func (r *Runner) Init(scripts []codersdk.WorkspaceAgentScript) error {
 	r.scripts = scripts
 	r.Logger.Info(r.cronCtx, "initializing agent scripts", slog.F("script_count", len(scripts)), slog.F("log_dir", r.LogDir))
 
+	err := r.Filesystem.MkdirAll(r.ScriptBinDir(), 0o700)
+	if err != nil {
+		return xerrors.Errorf("create script bin dir: %w", err)
+	}
+
 	for _, script := range scripts {
 		if script.Cron == "" {
 			continue
@@ -208,7 +227,18 @@ func (r *Runner) run(ctx context.Context, script codersdk.WorkspaceAgentScript)
 	if !filepath.IsAbs(logPath) {
 		logPath = filepath.Join(r.LogDir, logPath)
 	}
-	logger := r.Logger.With(slog.F("log_path", logPath))
+
+	scriptDataDir := filepath.Join(r.DataDir(), script.LogSourceID.String())
+	err := r.Filesystem.MkdirAll(scriptDataDir, 0o700)
+	if err != nil {
+		return xerrors.Errorf("%s script: create script temp dir: %w", scriptDataDir, err)
+	}
+
+	logger := r.Logger.With(
+		slog.F("log_source_id", script.LogSourceID),
+		slog.F("log_path", logPath),
+		slog.F("script_data_dir", scriptDataDir),
+	)
 	logger.Info(ctx, "running agent script", slog.F("script", script.Script))
 
 	fileWriter, err := r.Filesystem.OpenFile(logPath, os.O_CREATE|os.O_RDWR, 0o600)
@@ -238,6 +268,13 @@ func (r *Runner) run(ctx context.Context, script codersdk.WorkspaceAgentScript)
 	cmd.WaitDelay = 10 * time.Second
 	cmd.Cancel = cmdCancel(cmd)
 
+	// Expose env vars that can be used in the script for storing data
+	// and binaries. In the future, we may want to expose more env vars
+	// for the script to use, like CODER_SCRIPT_DATA_DIR for persistent
+	// storage.
+	cmd.Env = append(cmd.Env, "CODER_SCRIPT_DATA_DIR="+scriptDataDir)
+	cmd.Env = append(cmd.Env, "CODER_SCRIPT_BIN_DIR="+r.ScriptBinDir())
+
 	send, flushAndClose := agentsdk.LogsSender(script.LogSourceID, r.PatchLogs, logger)
 	// If ctx is canceled here (or in a writer below), we may be
 	// discarding logs, but that's okay because we're shutting down