I am wondering who else is actually using the test_slony_state script 
that we include in slony1-engine/tools

1) Is there anyone using the test_slony_state.pl script instead of the 
test_slony_state-dbi.pl.   The difference is one uses the old Pg perl 
driver versus the DBI::Pg driver.  Should we just remove the non DBI 
version from the source tree so we don't have to maintain both? (I can't 
think of why someone would be able to install the Pg driver, but not the 
DBI::Pg driver,  enlighten me if I am missing something)

2) How do people find the thresholds in that script, are people changing 
them to other values?  Should they be configured from the command line?

3) The script alerts if sl_log* grows bigger than 200000 tuples.  There 
was some thinking that this can lead to a lot of false positives. Are 
others finding that?  Do we really want to detect if the cleanup event 
is failing to run? This would have to do more with the age of sl_log_1 
versus the size.

4) The script outputs a bunch of stuff to stdout even when everything is 
okay.  Is this useful information or should it be silent unless an 
problem is detected? (Today it sends an email if a problem is found but 
always outputs stuff to stdout)

Steve