FYI: We came across the same issue, and were tempted to make this  
exact modification.

Chris: Can you double check the logic in this function. I don't ever  
see the log_txid get smaller (due to the cleanup thread no longer  
deleting entries from the table), so the condition is never met in  
2.0. If you manually truncate the log or force the cleanup thread to  
delete entries like it did in 1.2, it works, but this kills  
performance and defeats the purpose of trying to truncating the table.
____________________
John Stockdale
Sr. Systems Engineer
jstockdale at videosurf.com

> Good day seniores!
>
> >> Scott G. Miller wrote:
> >
> > After much fiddling around getting that script working (perl-pg is  
> not
> > readily available for RedHat, bugs in the script for perl 5.8,  
> etc), I get
> > no errors except complaining that the log sizes are large and that  
> a node
> > might be down.  Both nodes are up, and the number of lag events  
> ranges
> > between 0 and about 12, over about 10 seconds.  Everything appears  
> to be
> > well:
> >
>
> I have the same problem with Slony 2.0 and PG 8.3 and described it in
> slony1-general list in december.
> The only solution I've found is to manually edit logswitch_finish()
> function and replace:
> [code]
> if exists (select 1 from "_CKS_Cluster".sl_log_2 where log_origin =
> v_origin and log_txid < v_xmin limit 1) then
> [/code]
> to
> [code]
> if exists (select 1 from "_CKS_Cluster".sl_log_2 where log_origin =
> v_origin and log_txid >= v_xmin limit 1) then
> [/code]
>
> and also:
> [code]
> if exists (select 1 from "_CKS_Cluster".sl_log_1 where log_origin =
> v_origin and log_txid < v_xmin limit 1) then
> [/code]
> to
> [code]
> if exists (select 1 from "_CKS_Cluster".sl_log_1 where log_origin =
> v_origin and log_txid >= v_xmin limit 1) then
> [/code]
>
>
> Best regards, Nick Karih.