[Public] Deploying IronPDF in AWS Lambda: Understanding CustomDeploymentDirectory Behavior

Summary

When deploying IronPDF in an AWS Lambda function using Docker, developers may attempt to use a temporary directory like /tmp as a CustomDeploymentDirectory. However, this can result in misleading log messages or failed native binary lookups — even though the PDF rendering process still works.

This article clarifies:

• The role of CustomDeploymentDirectory

• Why using /tmp may not behave as expected

• When you actually need to set it

• Recommended workarounds and best practices

Issue Description

When setting IronPdf.Installation.CustomDeploymentDirectory = "/tmp" in AWS Lambda, IronPDF logs the following:

This may look like an error, but PDF rendering still proceeds normally — IronPDF simply falls back to using the default working directory (/var/task) where the binaries are already available.

Why This Happens

• CustomDeploymentDirectory serves two purposes:

  1. A directory to download and deploy native components (e.g., IronPdf.Native.Chrome.Linux)

  2. An alternative lookup path for runtime dependencies if not found in standard locations

• IronPDF prefers already-deployed assemblies and only falls back to the custom path if necessary

• In AWS Lambda, your Docker container's working directory is often already set to /var/task, where all required binaries reside

Hence, setting /tmp as a custom path is unnecessary unless you're using IronPdf.Slim deployments or relying on runtime NuGet downloads of the binaries.

Recommended Configuration

For most AWS Lambda + Docker deployments with IronPdf.Linux, the simplest and most reliable setup is:

IronPdf.Installation.AutomaticallyDownloadNativeBinaries = false;
IronPdf.Installation.LinuxAndDockerDependenciesAutoConfig = false;
IronPdf.Installation.ChromeGpuMode = IronPdf.Engines.Chrome.ChromeGpuModes.Disabled;
IronPdf.Installation.SingleProcess = true;

var awsTmpPath = @"/tmp/"; // AWS temporary storage path

// Configuration for IronPDF rendering

IronPdf.Installation.ChromeGpuMode = IronPdf.Engines.Chrome.ChromeGpuModes.Disabled;
IronPdf.Installation.DefaultRenderingEngine = IronPdf.Rendering.PdfRenderingEngine.Chrome;
Environment.SetEnvironmentVariable("TEMP", awsTmpPath, EnvironmentVariableTarget.Process);
Environment.SetEnvironmentVariable("TMP", awsTmpPath, EnvironmentVariableTarget.Process);
IronPdf.Installation.TempFolderPath = awsTmpPath;
IronPdf.Installation.CustomDeploymentDirectory = awsTmpPath;
IronPdf.Installation.LinuxAndDockerDependenciesAutoConfig = true;

🚫 When Not to Use CustomDeploymentDirectory

You do not need to set CustomDeploymentDirectory if:

• You are using IronPDF dependencies (e.g., IronPdf.Linux) during packaging of your project

• You're using full deployments and not downloading binaries at runtime

• The Lambda function can access everything from /var/task (default behavior)

✅ When You Should Use It

You should consider using CustomDeploymentDirectory = "/tmp" only if:

• You're using IronPdf.Slim without native binaries

• You rely on runtime NuGet downloads for components like IronPdf.Native.Chrome.Linux

• You want to store temporary deployments in the only writable AWS Lambda directory (/tmp)

Key Takeaways