[QFJ-294] "Initial Handshake failed" on ssl sessions with 1.3.1 Created: 01/Feb/08 Updated: 10/Jan/10 Resolved: 10/Jan/10 |
|
Status: | Closed |
Project: | QuickFIX/J |
Component/s: | None |
Affects Version/s: | 1.3.1 |
Fix Version/s: | Future Releases |
Type: | Bug | Priority: | Default |
Reporter: | Thomas Hügel | Assignee: | Unassigned |
Resolution: | Cannot Reproduce | Votes: | 0 |
Labels: | None |
Attachments: | ssl.zip |
Description |
Today I exchanged the quickfix/j library in my fixengine in our test environment with the 1.3.1 version(before 1.3.0). 19:30:01,865 | [SocketConnectorIoProcessor-3.0] | INFO initiator.InitiatorIoHandler | MINA session created: /10.252.212.72:50229 I switched back to the old library (1.3.0), restarted the fix engine and it works without any problems. the mina core and the ssl-filters havent changed, so it must be something in the quickfix/j core code. Configuration of the session: [SESSION] Regards |
Comments |
Comment by Steve Bate [ 02/Feb/08 ] |
Not much has changed in this part of the code. The only thing I'm seeing so far that might result in this type of error is a modified strategy for loading the keystore files. It should be backwards compatible. Do you see a warning with the text "keystore not found, using empty keystore" in your application log files? If so, that would indicate there is a file loading problem. |
Comment by Thomas Hügel [ 04/Feb/08 ] |
i tested again and i did not found any "keystore not found warnings". But the ssl functionality remains still broken. Switching back to 1.3.0 fixed all problems. |
Comment by Steve Bate [ 04/Feb/08 ] |
Very mysterious. The SecureSocketTest is passing so it's working to some extent. Can you turn on SSL debugging to see if it gives any more useful information? http://java.sun.com/j2se/1.4.2/docs/guide/security/jsse/JSSERefGuide.html#Debug In the meantime, I'll do some experimentation and see if I can reproduce the problem in more complex scenarios than the one being tested in SecureSocketTest. Steve |
Comment by Thomas Hügel [ 07/Aug/08 ] |
Steve, i tested on 1.3.2, but the SSL functionality is still broken. Have to stay at 1.3.0 . Thomas |
Comment by Steve Bate [ 08/Aug/08 ] |
Thanks for the report. I'd like to fix this issue but I've never been able to reproduce it. The test of the secure communications has continued to pass and I'm not sure what's different about what you're doing and what the test is doing. If you are able to create a variant of the secure socket unit test that fails it would greatly help me to diagnose and fix the problem. |