IIS Recycling and Virtual Memory Limit

In IIS (Internet Information Services), recycling refers to restarting the worker process (w3wp.exe) that handles web application requests. It's a built-in mechanism designed to improve the stability and performance of web applications by periodically refreshing the environment in which the application runs.

What is Recycling in IIS?

When a worker process is recycled, it is terminated and replaced by a new one. This can help:

• Free up memory leaks or excessive memory usage

• Reset unstable application states

• Apply configuration changes without restarting the server

Recycling can happen in various ways:

• Scheduled recycling (e.g., every 1740 minutes by default)

• Specific time recycling

• Request-based recycling (e.g., after a certain number of requests)

• Memory-based recycling (discussed next)

---

What is Private Memory Limit?

• Private Memory Limit defines the maximum amount of private memory (non-shared RAM) a worker process is allowed to use.

• If the limit is exceeded, IIS recycles the worker process to avoid potential performance degradation or memory exhaustion.

Example: If a site has a memory leak, the private memory keeps growing. When it hits the threshold (e.g., 500 MB), IIS recycles it to release memory.

---

What is Virtual Memory Limit?

• Virtual Memory Limit sets a cap on the total virtual memory (including both RAM and page file usage) a worker process can use.

• If the worker process exceeds this, it’s recycled.

This helps protect the server from apps consuming excessive memory, which could otherwise affect other applications or the whole server.

---

Key Differences

Metric	Type of Memory	When to Use
Private Memory Limit	RAM (non-shared)	To prevent apps with memory leaks hogging memory
Virtual Memory Limit	RAM + Page File (total VM)	For overall memory usage control

---

Things to Keep in Mind

• Frequent recycling can cause performance issues like application pool cold starts.

• Memory limits should be based on monitoring real-world usage, not arbitrary numbers.

• Set limits carefully, especially on high-traffic apps.