On Thu, May 17, 2012 at 11:57 AM, Vlad K. <v...@haronmedia.com> wrote: > PostgreSQL 9.0.7, libpq is part of the same version -devel package > PostgreSQL 9.1.3 (via pgsql yum repo), libpq is part of the same version > -devel package
9.1 (the one you have in production) also has leak-related fixes. They don't necessarily apply to your problem, they're somewhat esoteric, but it's worth noticing. Even the tiniest memory leak will fragment your process' heap, and cause it not to release those 200M. In those cases, especially when the leaks are one-time (like in the libpq case), it's convenient to force the leak to happen at load time. Precompiling queries in SQLA, to populate the various SQLA's compiler caches, doing some queries that cause libpq and psycopg2 to excercise (and thus to allocate whatever permanent data structures it needs to), all at load time, will help keep fragmentation to a minimum. Fragmentation is a complex issue, and both python and SQLA are quite prone to it. But it can be worked around. -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com. To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.
