Thanks for the input.

The truncate is what slony would send to a subscriber in order to clear out
the existing table before copying it.  I am not manually executing the
truncate.  This happens when you execute a subscribe set.  Basically, the
issue I need to understand is this: when I first replicate a table with data
in it, will the table truncate and copy as if it were a regular subscriber,
or do I need to do something special to the log shipped node.

Thinking back, I think you can ignore most of my original question and it
boils down to this: if I have a new table to replicate with data in it, will
the subscriber receive the initial truncate and copy, or do I need to "do
something?"

Thanks for your help!
Aaron


On 9/22/08 5:45 PM, "Jennifer Spencer" <jennifer.spencer at stanford.edu>
wrote:

> Hi Aaron,
> 
> It would be pretty easy to check the truncate behavior by running another slon
> process, with a test
> cluster and table, and trying it out.  I myself don't know if "truncate" is
> something Slony explicitly
> supports (and I am guessing not).
> 
> I'd think about it this way: the log shipping process is only another "slon"
> process (with an archive
> flag -a to save out all the commands issued by the slon).  If your existing
> slon process to your
> regular subscribers won't do what you want (truncate), then your slon -a (log
> shipping) won't do it
> either.  If it will, then the slon -a should also.  See what I mean?
> 
> I assume from your mail that you are going to have your subscribers issue a
> "truncate" on their end,
> and do some sort of external-to-Slony process to re-replicate the table in
> question?  Treat your
> log-shipping subscriber the same as your regular subscribers.  Just before you
> truncate and munge your
> data, turn off your slon -a process.  You should note the log number in
> _yourclustername.sl_archive_tracking when you turn off your slon.  Make sure
> to send that number to
> your remote log subscriber.  He'll need to ingest the last CORRECT pre-munge
> log (that number you're
> going to send him), and then turn off his log-ingest process after that.  Then
> he would do the 
> truncate on the data table, receive and ingest a new set of munged data from
> you (via your copy
> out/his copy in or another re-replication process you had in mind).  You then
> restart your log-ship
> slon -a process, and then he'll restart his log-ingest process.  Just make
> sure not to make other
> changes to other tables when your slon -a is inactive.
> 
> Hope this helps.
> -Jennifer
> 
> Aaron Brown wrote:
>> List,
>> 
>> I have a large table (16 million rows), that I need to pull out of
>> replication, do some munging, and then rereplicate (in a new set) as part of
>> a release.  To explain better, the process will be to drop the table (which
>> is in set 2) from replication, run a large stored procedure that munges the
>> data, then rereplicate the table by adding it to a new set (set 12).  The
>> subscribers obviously truncate the table and rereplicate the entire table.
>> This process works fine.
>> 
>> The catch here is that we have a remote log shipped node off of one of the
>> subscribers.  I know the schema changes themselves do not propagate.  What
>> I¹m unclear on is whether or not the log shipping archive files will contain
>> the TRUNCATE and COPY statements to recreate this table, or if I will have
>> to do something (like recreate the remote node from scratch) to make that
>> happen? 
>> 
>> Thanks ­ any help you can provide is appreciated.  If you need more
>> specifics or configs, please let me know.
>> 
>> Aaron
>> 
>> 
>> 
>> ------------------------------------------------------------------------
>> 
>> _______________________________________________
>> Slony1-general mailing list
>> Slony1-general at lists.slony.info
>> http://lists.slony.info/mailman/listinfo/slony1-general


-------------------------------------------------------
Aaron Brown, Systems Engineer
BzzAgent, Inc. | www.bzzagent.com
abrown at bzzagent.com | 617.451.2280
-------------------------------------------------------