Hi.  Sorry for my late reply - I appreciate your answers.

[I see a 1:89 ratio between SQL UPDATE's and (INSERT's + DELETE's) and
wanted to find out why.]

Dear Christopher:   I've got 21 sequences in my replication set.
Thank you for the vote of confidence in 2.0.5.  Thanks for the data
point that Slony does do SQL UPDATE's.

Dear Steve:  Yes, it sounds like I've run into bug #126.    I don't
really have anything to add to the bug ticket (besides "me too").  If
I can help gather any additional information, please let me know
what...   I'm quite interested in solving this --  we're a small shop
and any time a system is not self-healing, it can mean less sleep for
me.

Dear Vick:  I do understand running slony replication will have SOME overhead.
I just want to understand difference between 1.2 and 2.  If there is a
big difference,
it's a clear argument to move to 2.  Looks like I should do some
testing in our test lab.
And Vick, you asked:

On Thu, Oct 7, 2010 at 1:01 PM, Vick Khera <vivek at khera.org> wrote:
> How many rows are affected per update?  Slony replicates each row
> specifically, not the queries themselves.  So if you update 10000
> rows, there will be 10000 records for slony to copy over to the other
> end.

I have no idea.  I didn't write the queries, but I am (by default) the
DBA.  How
could I tell how many rows are affected per update?  (Feel free to RTFM me
in the write direction.)

Thanks very much,
Aleksey