I don't know what the relative performance of your systems are, but it sounds like you ran a single delete command that deleted a very large number of rows. Recall that slony is a data (not statement) base replication system. This means that you have generated a very large number of entries in the sl_log table on the origin, one for each row of data you deleted. These must be transfered to the replica which takes time, and then applied in a single transaction, which also takes some time. If the replica database appears to be busy grinding away then you can fairly reasonably assume that it's busy replicating the huge SYNC event you created. If you feel a strong need to verify that it's doing what it should be, you can either increase the slon log output level and see what it's doing that way, or use any of the traditional ways to snoop into the activities of the slon database connection.
Consider also that you have caused large changes in the tables and may not have vacuumed / analyzed tables sufficiently which is in turn hurting your performance. Andrew On 3/30/07, Andrew Edson <[EMAIL PROTECTED]> wrote: > I am attempting to work with a test database structure; I have two test > servers, both of which are running SuSE 9.3, PostgreSQL 8.1.3, and Slony > 1.1.5. > > I created the databases yesterday evening, giving them both the same schema, > and allowed the first one to fill overnight (I'm trying to work with data > from another, production, db, my boss hopes to set up replication on it > across our network in the near future). This morning I attempted to > replicate the data from the first to the second, following the instructions > in the Slony-1 Replication Tutorial. The data appears to have replicated at > least once; the formerly empty db now has the same data in it as the filled > one had. > > The next step I tried, however, was to run a delete statement. The end > result was to remove almost eight million records from a table on the > primary db. (I expected this number; there were about fifteen million > records on the table to begin with, and I used a fairly common value in my > filter.) The two slon instances appear to still be running, but they do not > appear to be communicating any longer; it has been two hours since the > delete ended, and I am not seeing any sign that the results of the delete > are being replicated. I've replicated delete statements on smaller db's > while trying to teach myself how to set up replication, so I know it can be > done. > > Does anyone have any advice as to how I can check to ensure that my delete > is getting replicated? I admit that the original delete statement may have > been a bit overboard in its scope, but I'd like to make sure I've got > everything working. > > Thank you for your consideration. > > ________________________________ > 8:00? 8:25? 8:40? Find a flick in no time > with theYahoo! Search movie showtime shortcut. > > > _______________________________________________ > Slony1-general mailing list > [email protected] > http://gborg.postgresql.org/mailman/listinfo/slony1-general > > _______________________________________________ Slony1-general mailing list [email protected] http://gborg.postgresql.org/mailman/listinfo/slony1-general
