[pgpool-general: 4320] Re: How to test pgpool?

[Tatsuo Ishii]

> I would appreciate some guidance if somebody can provide it.  The documentation tells me all the pieces but I still don’t get how they all fit together.

I do not fully understand your questions but...

> I have a 3 node setup. 1 master, 1 slave, pgpool.  I have pgpool running (9.4) and current using a master/slave setup.  How exactly does one test pgpool?
>
> 1. How exactly does it know weather to send something to master or slave node?  I’m aware of the load balancing feature.  What actually happens if I start a transaction and have 5 reads, 1 write, 4 more reads (in that order)?

Pgpool has a built in SQL parser inside. By using it, pgpool can know
whether it gets update the database. If so, pgpool will route the
query to the master (primary) PostgreSQL. If the query does not update
the database, pgpool sends it to either the master or one of the
standbys depending on the setting of configurations related to load
balancing.

You could know which database server pgpool routes queries by using
log_per_node_statement.

> 2. Does pgpool have a way to pass credentials along to the db nodes?  I’m interested in the write being recorded for log / tracking issue.  Can I have it setup to where the user always authenticates or does PGPool use a singular user account to do reads / writes?  A scenario would be:
> 	A. user H hits pgpool and does a select, followed b an update.
>   	B. user J hit’s pgpool and does an update.
> 
> It’s important that I track both users for auditing purposes.

I am not sure what you exactly mean by "credentials" here but you
could track which user sends which query by using
log_per_node_statement.

> Thanks in advance!
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general