History: The history of this error, which has mostly gone unsolved, dates back to Windows 2000.
Platforms affected: Windows Server 2008 R2, Server 2008, Server 2003 R2, Server 2003, Server 2000 (both 32-bit and 64-bit installs are affected).
Error Messages
Event ID: 7024 The Routing and Remote Access service terminated with service-specific error 2 (0x2).
Event ID: 7024 The Routing and Remote Access service terminated with service-specific error 31 (0x1F).
Event ID: 7024 The Routing and Remote Access service terminated with service-specific error 20205 (0x4EED).
Event ID: 7024 The Routing and Remote Access service terminated with service-specific error 193 (0xC1).
Event ID: 20103 Unable to load C:\WINDOWS\System32\iprtrmgr.dll
. (32-bit installs).Event ID: 20103 Unable to load C:\WINDOWS\SysWOW64\iprtrmgr.dll
. (64-bit installs).
(This is KCotreau's original answer, but he included it in the question. Re-posting as an actual answer)
Cause: There are two basic causes for this error.
Related causes
This has happened to me on a couple of Dell Windows 2003 R2 64-bit servers that had Broadcom NetXtreme II adapters. My problem was the second cause above, TCP/IP corruption, which I believe that happens for some reason when you install the driver for the Broadcom adapter. It certainly may happen with other adapters, but there were definitely a high number of unsolved cases on the Internet with various Broadcom adapters.
Additional factors
This happened even on clean installs as I tested using both the Dell-specific Windows Server 2003 R2 media, and Microsoft media downloaded from the Volume License site. It happens with either media.
Troubleshooting that did not work for me, or most people on the Internet:
NETSH INT IP RESET C:\reset.log
orNETSH RESET WINSOCK
.NETSH WINSOCK RESET
.devmgr_show_nonpresent_devices=1
and then showing hidden devices in Device Manager (I had none).C:\Windows\System32\ias \dnary.mdb
, andias.mdb
files and restarting RRAS. Those files are located in theC:\Windows\SysWOW64
directory on 64-bit systems. This was in KB840696 http://support.microsoft.com/kb/840686.Solutions
HKEY_LOCAL_MACHINE\System\currentcontrolset\services\remoteaccess\routermanagers\IPV6
, which you will remove after you back it up by exporting it. Removing this key was an easy solution that widely helped many people on the Internet. If you still want to use the TCP/IPv6 protocol you may have to do more. Since I did not apply to my server, and I could not test it, you may still have IPv6 corruption, and may want to troubleshoot by removing and reinstalling the TCP/IPv6, akin to the solution below. The above solution may just be masking potential corruption by avoiding the issue.If you are not running IPv6, the chances are that you have TCP/IPv4 corruption, and the solution is to reinstall it. If you have never noticed, if you try to uninstall TCP/IP, it is grayed out. To get around that, I followed KB 325356: http://support.microsoft.com/kb/325356 . That says it is for a domain controller, but works on member servers also. The steps are:
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Winsock
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Winsock2
c:\windows\inf
, and then click OK.c:\windows\inf
, and then click OK.Configure and Enable Routing and Remote Access. At this point, your RRAS should start.
Related problem and solution:
When I would enter a static IP address, although it would hold the static, the properties would reset back to "Obtain an IP address automatically". Navigate to
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Network
and delete the Config key. Reenter your static settings, and they should hold, and rebuild that Config key.I hope this solves this difficult problem for many of you.
I know this is a hecking old post, but the same unhelpful error comes up even in Server 2016.
Luckily, for me the cause was mismatched certificates between IIS and RRAS.
The fix was to use the same certificate for the default IIS site as I was using in RRAS.
RRAS on Server 2019 and 7024 or 20269/20153 events from RemoteAccess, after 4405 event from NPS
I changed the LogFile path in NPS -> Accounting. Afterwards RRAS started.
The solution in my case was quite simple: