http://www.slony.info/bugzilla/show_bug.cgi?id=252

--- Comment #2 from Mark Summers <womagrid at gmail.com> 2011-12-06 09:01:13 PST ---
(In reply to comment #1)
> I don't view this as a bug at all.
> 
> Although it is possible that the paths to/from the new node use the same
> conninfo string as the ones for the node that is cloned, we do NOT make that
> assumption. The cloning process requires explicit STORE PATH commands to
> provide this information.
> 
> I have created a patch to make the temporary conninfo string '<event pending>'
> so that it is consistent with the rest of the code.
> 
> https://github.com/wieck/slony1-engine/commit/a8273c700dfdeefa1c82944746384df6caa25236

Thank you for looking at this.

The current behavior causes ERROR log entries, which are potentially confusing.
 Your proposed change should address this.

I don't expect the cloning process to figure out or assume the value of
conninfo and I always use STORE PATH.  The issue is probably that I updated the
minimum paths required for replication and did not update the other paths that
were created historically by meta scripts, thereby leaving some 'Event pending'
sl_path entries behind.

So, this issue can be considered resolved.

As an aside, it seems to me that the cloning process, while extremely useful,
is rendered quite inaccessible by the absence of a worked example, including
important points such as when it is necessary to suspend writes to the database
and acknowledging the requirement for explicit STORE PATH.

These things might sound obvious but I was only able to infer the correct
sequence of events by considering the stored procedure implementation of the
prepare and finish functions, and Bernd Helmle's blog posting remains the most
complete documentation I have found.

-- 
Configure bugmail: http://www.slony.info/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.