[pgpool-general: 733] nonlogged-tables / question
phofstetter at sensational.ch
Sun Jul 15 01:24:21 JST 2012
in our application, we're using a unlogged (9.1 feature) table which
we join against. The table is used as a cache which can be easily (but
at some perfomance cost) rebuilt from the real data, so we chose to
make the table unlogged in order to get the maximum insert
My question now is how pgpool handles this when used with streaming
replication, as you can't query for unlogged tables on a hot standby
Some naïve testing (sending a read query joinin against that table
through pgpool over and over) has never shown a problem, so I *think*
pgpool knows about unlogged tables, but I'd rather have confirmation
on this :-)
So the question is: Does pgpool know about unlogged tables? Does it
send queries joining against unlogged table to master?
If not or not always, how can I convince it to do so?
As I said, from naïve testing, I would assume that pgpool does it
right - I would just love to have confirmation.
More information about the pgpool-general