What causes "Failed to compress %s: %s" in hashicorp/packer?

Check free disk space at the output path (df -h) and clean up

Failed to compress %s: %s

Error message

Failed to compress %s: %s

What it means

The io.Copy from the artifact's source file into the compression writer failed during Packer's compress post-processor. Reading the source or writing/compressing into the output stream returned an error after the file was successfully opened, so the build aborted mid-compression.

Source

Thrown at post-processor/compress/post-processor.go:243

		// into our compressor.
		if len(artifact.Files()) != 1 {
			return nil, false, false, fmt.Errorf(
				"Can only have 1 input file when not using tar/zip. Found %d "+
					"files: %v", len(artifact.Files()), artifact.Files())
		}
		archiveFile := artifact.Files()[0]
		ui.Say(fmt.Sprintf("Archiving %s with %s", archiveFile, compression))

		source, err := os.Open(archiveFile)
		if err != nil {
			return nil, false, false, fmt.Errorf(
				"Failed to open source file %s for reading: %s",
				archiveFile, err)
		}
		defer source.Close()

		if _, err = io.Copy(output, source); err != nil {
			return nil, false, false, fmt.Errorf("Failed to compress %s: %s",
				archiveFile, err)
		}
	}

	ui.Say(fmt.Sprintf("Archive %s completed", target))

	return newArtifact, false, false, nil
}

func (config *Config) detectFromFilename() {
	var result [][]string

	extensions := map[string]string{
		"tar":   "tar",
		"zip":   "zip",
		"gz":    "pgzip",
		"lz4":   "lz4",
		"bgzf":  "bgzf",

View on GitHub (pinned to eb36e3c3e4)

Solutions

Check free disk space at the output path (df -h) and clean up
Look at the wrapped inner error to distinguish read vs write failure
Move output_directory to larger/local storage and retry
Retry the build; if persistent, test the source file integrity (e.g. fsck or re-download)

Defensive patterns

Strategy: retry

Validate before calling

// ensure adequate disk space before compressing
if fi, err := os.Stat(outputDir); err == nil {
    _ = fi
}
// check at least 2x the source size is free on the output filesystem
if st, err := os.Stat(sourcePath); err == nil {
    if free, _ := diskFree(outputDir); free < 2*st.Size() {
        return fmt.Errorf("insufficient disk space to compress %s", sourcePath)
    }
}

Try / catch

// Go: retry once on transient copy failure
artifact, _, _, err := pp.PostProcess(ctx, ui, artifact)
if err != nil && strings.Contains(err.Error(), "Failed to compress") {
    time.Sleep(2 * time.Second)
    artifact, _, _, err = pp.PostProcess(ctx, ui, artifact)
}
return err

Prevention

Pre-check free disk space (>= 2x input size) on the output volume
Avoid concurrent jobs writing to the same output directory
Monitor the source file isn't modified during the build
Use local disks for output rather than unreliable network storage

When it happens

Trigger: The default (single-file, non tar/zip) branch runs io.Copy(output, source) and it returns an error — typically ENOSPC writing the output archive, an I/O error on the output file, or a read failure on the source mid-stream.

Common situations: Disk filling up while compressing a large disk image; output filesystem hitting a quota; source file truncated/changed concurrently; failing disk or network storage holding the output path.

Related errors

AI-assisted analysis of hashicorp/packer@eb36e3c3e4 (2026-09-05). Data as JSON: /api/errors/ad39064932764a89. Report an issue: GitHub.