Bugzilla – Bug 1241
Intermittent failures in Big_Twoways
Last modified: 2002-06-27 09:30:44
You need to log in before you can comment on or make changes to this bug.
The Big_Twoways test fails regularly on the nightly regression builds with a TRANSIENT exception. Apparently the failure is due to multiple simultaneous connects() from all the peers taking to each other. Probably related to bug 189. I have a fix (that will attach promptly.)
Created an attachment (id=129) [details] Patches for bug 1241
Could you please go ahead and check this patch in? It will be useful for my testing. Thanks
I attached the proposed fixes. Basically the test forces the ORB to open the connections before running the main event loop. I ran the test overnight on a dual PII @ 1Ghz and had no failures.
I am applying the patches now. I can test it on my local win32 box but not on Linux or solaris now since the whole network is down!
Fixed, please see the following ChangeLog entry for details: Thu Jun 27 10:28:29 2002 Carlos O'Ryan <coryan@atdesk.com> similar fixes may be required in other tests (Big_Oneways and LongUpcalls seem to suffer the same problem)