[pgpool-general: 828] Documentation Questions
mc69921 at matthewc.net
Tue Jul 31 13:03:21 JST 2012
Up until now, I have only considered lighter weight connection poolers, but with 3.2's more transparent memory query cache,
pgpool-II has suddenly become my primary interest for new deployments. It looks like it it should work well with all open source
projects that only use basic DML (no triggers, SPs, or update'able views). I am specifically thinking about MediaWiki and Drupal.
I was reading through the documentation for pgpool-II-3.2.0-RC2 and found a 3 things:
1) Documentation: "On memory cache saves pair of SELECT statements (with its Bind parameters if the SELECT is an extended query)."
I am not familiar with its design, but would think that the actual pair is the select statement (with bind parameters if applicable)
and the accompanying result set. The sentence implies that there are two SELECT statements per cache entry. I do not see what a
second SELECT statement would be for.
2) Does black_memqcache_table_list also blacklist tables if the memqcache_method is shmem? The name implies that it only operates
with the memqcache method. In which case, the shmem method appears to not have table blacklist functionality. If that is the case,
maybe it is worth mentioning that memqcache has this exclusive advantage in the documentation.
3) Documentation under Restrictions \ Query Cache:
"Currently, the query cache must be deleted manually. pgpool-II does not invalidate old query cache automatically when the data is
This sounds like it only applies to versions 3.1 and earlier. Do you really have to manually invalidate for version 3.2 and later?
I noticed that you use MediaWiki with PostgreSQL 9.0 at http://pgpool.net/mediawiki/. Assuming that the query cache works with
MediaWiki, are you currently testing pgpool-II-3.2.0-RC2 with query caching enabled for your MediaWiki install?
More information about the pgpool-general