|
Bugzilla – Full Text Bug Listing |
| Summary: | [System.Web.Services] SOAP-Problems | ||
|---|---|---|---|
| Product: | [Mono] Mono: Class Libraries | Reporter: | Gen2 <spielc> |
| Component: | Sys.Web.Services | Assignee: | Atsushi Enomoto <atsushieno> |
| Status: | RESOLVED FIXED | QA Contact: | Mono Bugs <mono-bugs> |
| Severity: | Normal | ||
| Priority: | P3 - Medium | CC: | anmar, gert.driesen, informatique.internet |
| Version: | 1.2 | ||
| Target Milestone: | --- | ||
| Hardware: | Other | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Found By: | --- | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Bug Depends on: | 371859 | ||
| Bug Blocks: | |||
| Attachments: |
c#-ws-proxy for http://www.atomic-x.com/xmlservices/GuitarManufWS.asmx?wsdl
c#-proxy for my sample ws asmx-file of my web-service implementation of this simple web-service All in one test case from the above attachments (use this one to reproduce) |
||
|
Description
Thomas Wiest
2007-09-15 20:13:26 UTC
xsp2 does not cause that ReadDone() error on web browser (DefaultWsdlHelpGenerator.aspx) on Linux. On windows it has been a long issue though (I've never seen it worked). Marek, any idea on this general issue? Creating, compiling and running the simple web service proxy works fine here on Windows. However, the Test Form indeed results in the following error using both XSP and XSP2: "Error getting response stream (ReadDone2): ReceiveFailure" XSP(2) logs the following exception: Peer unexpectedly closed the connection on write. Closing our end. System.IO.IOException: Write failure ---> System.Net.Sockets.SocketException: An established connection was aborted in your host machine. at System.Net.Sockets.Socket.Send (System.Byte[] buf, Int32 offset, Int32 size , SocketFlags flags) [0x00000] at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x00000] --- End of inner exception stack trace --- at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x00000] at Mono.WebServer.XSPWorker.Write (System.Byte[] buffer, Int32 position, Int32 size) [0x00000] I experienced this problem on our intranet (10x/ days) but i can't reproduce it on my dev PC... But this bug is very annoying... This looks like something we came into a few months ago (at least the part about axis). Problem was that axis client makes requests with chunked encoding and xsp2 does not support it. With mod_mono is even worse, as apache automatically decodes chunked encoding requests, but is doesn't provide the content-length expected by mod-mono-server. I have a couple dirty patched bitrotting somewhere that addressed this, though I never got around to cleanup them. Hubert, can you describe what is the web server that hosts the web service that we are talking to? Hubert case is almost ignorable or different from this issue because he says that he cannot reproduce the issue (while everyone else can). He should file different bug report. Adding bug dependency (now we simply cannot access to *.asmx). (In reply to comment #5 from Miguel de Icaza) > Hubert, can you describe what is the web server that hosts the web service that > we are talking to? > The issue appears now 3-4 times/ days... My webserver is running asp.net over an nfs share (the aspx and the dll are on an nfs share in order to do load balancing). The nfs server is serving a SAN disk schema of our architecture <ASP WebServer1><--> <WebService WebServer1> <ASP WebServer2 <--> <WebService WebServer2> | | |------------------------- <Nfs Server> | <SAN> There's something strange, two weeks ago, the write cache of our SAN has been disabled, write performance on our nfs was awful... and the problems happened 60-70 times/days... It seems there's a link between nfs and this problem (that's why i can't reproduce it on my dev machine which is not on an nfs share Looks like the link (http://www.atomic-x.com/xmlservices/GuitarManufWS.asmx?wsdl) is no longer valid. Is this bug still reproduceable without that link? Created attachment 220559 [details]
All in one test case from the above attachments (use this one to reproduce)
Using the all in one test case from comment 9, I am unable to reproduce this bug on Mono SVN r105066 on SLED 10 SP2 i586. I'm closing this bug as I think it's been fixed. Also, from comment 6 it seems that Hubert should file a new bug for his comments on this bug. Gen2, if you can still reproduce this, please re-open this bug. |