Affects Version/s: 3.0.0, 3.0.1, 3.0.2, 3.0.3, 3.0.4, 3.1.0
Fix Version/s: 184.108.40.206
Windows Server 2016 on VMWare
Our website is running mostly Classic ASP, I have set useHeaders="true" useVariables="false"
Web Server:IIS 9 (Windows 2016)
A few months ago we started using the Shibboleth Service Provider v220.127.116.11 on IIS on Windows Server 2016 (with all recent Windows updates)
It can be running fine for days but then sometimes w3wp.exe crashes (a few times on a day) with no apparent cause.
Windows Error Reporting shows:
The 0xe06d7363 translates to EXCEPTION_UNCAUGHT_CXX_EXCEPTION
Next step is to catch a minidump. The Call Stack is shown below
Iis7_shib.dll is very high in the call-stack with IIS7Request::throwError at the top.
It is called by IIS7Request::getRequestBody because a call to m_request->ReadEntityBody fails.
First of all: I am not sure Shibboleth is *causing* the crash, maybe it is just collateral damage. But it may help in finding the root cause
- A message goes into the IOException. Is there any way to retrieve that message (some logging perhaps? I could not find it in shibd.log or any other log)
--> In the minidump I can see the error is ERROR_NETNAME_DELETED
- ReadEntityBody sound like it is receiving data. Is it fair to generate an (unhandled) exception for ‘just’ a network error?
- If it is only collateral damage: is there a way to catch the exception?
I do see the <InProcess> element supports a (disadvised) catchAll but if I read the sourcecode correctly it may not apply to the IOException
(callstack and shibd.log follows)
Also: this is what shibd.log shows during the crash: