Hi all;
I have a 3 node cluster as follows:
SLONY 2.0.7 on all three nodes, each install of slony was built on each
node via source
nodes 1 & 2 = postgres 9.1.0
node3 = postgres 9.0.4
I did this:
1) I used pgbench -i to initialize the master database before I setup &
started replication
2) I created a new column as a primary key on the pgbench_history table
like so
alter table pgbench_history add column pgb_hist_id serial;
alter table pgbench_history add primary key (pgb_hist_id);
4) then I ran pgbench like this:
pgbench -t 800 rep_bench_node1
5) after replication was started both slaves sync'd up fine
at this point all 3 nodes have 800 rows in the pgbench_history table
6) then I run pgbench again, reducing the number of tx thus reducing the
overall number of rows in the pgbench_history table
$ pgbench -t 300 rep_bench_node1
starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 1
query mode: simple
number of clients: 1
number of threads: 1
number of transactions per client: 300
number of transactions actually processed: 300/300
tps = 265.106193 (including connections establishing)
tps = 285.114730 (excluding connections establishing)
7) the master (node1) now has 300 rows in the pgbench_history table
however both slaves now have 1100 rows (it didn't do the delete on
the slaves)
at this point each subsequent run of "pgbench -t 300
rep_bench_node1" replaces the rows in node1 but appends another 300 rows
to each slave.
Thoughts?
Am I doing something wrong?
Thanks in advance
--
---------------------------------------------
Kevin Kempter - Constent State
A PostgreSQL Professional Services Company
www.consistentstate.com
---------------------------------------------
_______________________________________________
Slony1-general mailing list
[email protected]
http://lists.slony.info/mailman/listinfo/slony1-general
