David Janssens wrote: > Hello, > > I would like to know why sync events are generated in my cluster every > 10s even when nothing changes? 1. Sequences are polled, so this keeps them up to date 2. The frequency is configurable using the "-s" and "-t" options so you may change the timing two ways > As far as i know, this causes some useless work (growing sl_event > table, network traffic between nodes, etc.) > Also i would like to know if it's possible to replicate tables as soon > as they are modified, instead of having to wait 10s. > You might use -s 1000 so that it checks every 1000ms rather than every 10000... Setting it extremely low won't be overly good - that'll make Slony-I poll for work *exceedingly* frequently, and each SYNC has some cost, so pushing through many times as many SYNCs is liable to be costly.
You might also increase from the default "-t 60000" to something higher so that if no updates are coming in, SYNCs are generated less often. _______________________________________________ Slony1-general mailing list [email protected] http://gborg.postgresql.org/mailman/listinfo/slony1-general
