Christopher Browne cbbrowne at ca.afilias.info
Mon Sep 24 14:29:28 PDT 2007
JP Fletcher <jpfletch at ca.afilias.info> writes:

> Hi,
>
>
> I'm able to reproduce this condition repeatedly.  The scenario is as
> follows:
>
> Set1 has origin on node8143, which is data provider to node8141, which
> in turn is provider to node8194 :
>
> 8143 --> 8141 --> 8194
>
>
> I add set2, which has origin on node8143.  I then subscribe node8194
> to node8143 directly.
>                                  This breaks node8194 immediately.

I have added a test case that exercises this kind of scenario to CVS HEAD:
<http://lists.slony.info/pipermail/slony1-commit/2007-September/002012.html>
<http://lists.slony.info/pipermail/slony1-commit/2007-September/002013.html>

The test is called "testmultipaths."

We'll be able to readily measure when the problem is repaired.

I don't believe that this kind of case has ever worked, but it's
definitely a bad thing that it doesn't.
-- 
select 'cbbrowne' || '@' || 'cbbrowne.com';
http://www3.sympatico.ca/cbbrowne/advocacy.html
Overheard in the mall: "What's with your sister and her spitting fetish?" 


More information about the Slony1-general mailing list