Jan,

>> Artificially delaying the replication process for purposes of point in
>> time recovery was considered in the original design of the Slony-I
>> system, but the mechanism was actually never implemented. The idea was
>> not quite how you describe it above, where slon would wait until some
>> sort of command kicks off the catch up process, but rather that slon
>> would ignore events unless they are at least X minutes old. This could
>> relatively simple be implemented in the query that looks for new events
>> in the remote_listen thread, by adding a qualification that suppresses
>> those "too new" events.

>> I will see if we can easily add the feature to the upcoming 2.0 release
>> of Slony-I.

Is it your intention to implement (or investigate the possibility of
implementing) delayed replication as I've described it, or to implement the
"ignore unless x minutes old" model, or both?

Thanks in advance,
Peter

>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20071211/=
af1bd689/attachment.htm