Hello,

We are currently running a Postgres 8.1.3 database on an Ubuntu 5.10 
server and recently implemented slony to replicate it.

We're encountering an issue when including the sequences in our db 
schema in the replication set. When replicating only our tables (about 
100 tables in total), performance and replication is smooth. However, 
when including the sequences in the replication process (about 14 
sequences), this system load on our database server skyrockets after a 
couple of hours and the DB becomes unresponsive to any communication 
from our application. Without the sequences included in the set, the 
load average hovers around the 0.2 up to the 2.0 range when being stress 
tested, but when including them, the load rises to 10 and more after a 
few hours of normal operation, at which point the server eventually 
becomes unresponsive.

While I realize it is warned that replicating sequences incurs a 
performance hit, I can't imagine such an extreme effect to be expected. 
Are there any configuration options you can suggest to us to explore and 
experiment with to deal with this before reporting anything on the Slony 
bug tracker?

Thanks,

Costa Galanis