Fix RPC_NT_ENTRY_NOT_FOUND 0XC002003E on Windows

When you see this error

You're working on a Windows Server 2019 or 2022 box, maybe running a legacy app that calls a DCOM component or a PowerShell script that queries WMI remotely. Suddenly you get RPC_NT_ENTRY_NOT_FOUND (0XC002003E). The error message says "The entry was not found". This usually happens after you reboot the server, restart a service, or change a firewall rule. The app worked fine yesterday. Today it doesn't.

Root cause

What's actually happening here is your client (local or remote) tries to connect to an RPC endpoint — a specific port and interface that a service registered with the RPC Endpoint Mapper (EPM) running on port 135. The EPM keeps a dynamic table of which service listens on which port. When the error fires, the EPM answered the request but said, "I have no record of that interface UUID."

The reason for that is almost always one of three things:

The target service isn't running. If the service never registered its endpoint, the EPM has nothing to give you.
The service registered but crashed or stopped after registration. Windows services can start, register endpoints, then die silently — leaving stale entries or none at all.
The RPC subsystem itself is broken. This is rare but happens if something corrupts the RPC runtime (malware, failed update, disk corruption).

Unlike RPC_S_SERVER_UNAVAILABLE (which means the machine didn't answer at all), this code means the machine answered but the endpoint list is empty for that interface.

Fix it — step by step

Check if the RPC Endpoint Mapper is running. Open Services.msc. Look for Remote Procedure Call (RPC). Its status should be Running. If not, start it. But if it stopped, something else is wrong — check the System event log for earlier errors about RPC crashing.
Identify the interface UUID that failed. Go to Event Viewer → Windows Logs → System. Filter by Event ID 10005 (DCOM errors) or 7031 (service crash). The error message often includes the failing interface UUID like {000001A0-0000-0000-C000-000000000046}. Write that down.
Find which service owns that UUID. Open an elevated PowerShell prompt. Run:
```
Get-WmiObject -Class Win32_ClassicCOMClassSetting | Where-Object { $_.ProgId -match 'your component name' }
```
If you don't know the component, search online for the UUID. Microsoft publishes a list of well-known UUIDs. Often it's IID_ICOMObject from a Microsoft service like Windows Update or a third-party app.
Restart the owning service. Once you know the service, do a hard restart:
```
net stop <service_name> && net start <service_name>
```
Then check if the error goes away. If it's a DCOM app, restart the DCOM Server Process Launcher service too — but careful, restarting that kills all COM objects temporarily.
Check firewall rules for port 135 and dynamic RPC ports. The EPM uses port 135. After it answers, the client connects to a random high port (49152-65535 on modern Windows). If a firewall blocks these dynamic ports, the client hangs but the error logged will still be this one. To test, temporarily disable the firewall. If the error stops, you need to open the RPC dynamic port range.
Repair the RPC runtime. If nothing above works, run sfc /scannow from an elevated command prompt. If it finds corruption, reboot and test again. In extreme cases, run dism /online /cleanup-image /restorehealth to fix the component store.

If it still fails

Check the Application event log for Event ID 7034 — that means a service crashed unexpectedly. Also check the System log for Event ID 1000 (application error) from the RPCSS process. If you see repeated crashes, the service itself is buggy. Try reinstalling the app or applying the latest Windows cumulative update. On Server Core installations, you might need to install the RSAT-RDS-Server feature if the missing endpoint belongs to Remote Desktop Services. The real fix is knowing which UUID failed — once you have that, the answer is usually obvious.