On Wed, 2010-01-06 at 16:39 +0100, Adrian von Bidder wrote: > Howdy! > > I'm currently in a MySQL -> PostgreSQL migration project (Go, go, go, ... > shall I cc: slashdot, too? ;-) > > Part of this is in embedded context, where a (diskless) embedded computer > runs from flash. Since we don't want to stress the flash too much, the db > is actually loaded from a dump at boot and run off a ramdisk (and dumped > back to flash periodically.) > > So all the nice database crash-safety that pg offers goes down the drain... > > With this in mind: what can we tune to minimize diskspace (RAM-disk space) > usage? And, possibly, RAM usage as well (disk access is, after all, almost > as fast as RAM access in this scenario) > > With our test dump, the db (after import) is ca. 300M on disk, ca. half in > WAL files (pg_xlog.) If I could mostly get rid of the WAL (keep it to a > bare minimum and run pg without fsync, something like that), the remaining > 160 to 180M would be ok. > > Next thing would be: are there tunables (compile time?) to minimize the > actual size of the rows? > > Your input is valued, thanks in advance. Wow, here are some things but honestly Pg may not be the DB you are looking for. You can set checkpoint_segments to 3. That will use 48 megs at any given time but will force a checkpoint (likely quite a few) during your load. This will slow the load down. As far as rows... there isn't really anything you can do. The data and its representation is the data. I thoroughly applaud your willingness to move to Pg but in this case something like SQLite or even Ingres may be a better solution for you. Sincerely, Joshua D. Drake -- PostgreSQL.org Major Contributor Command Prompt, Inc: http://www.commandprompt.com/ - 503.667.4564 Consulting, Training, Support, Custom Development, Engineering Respect is earned, not gained through arbitrary and repetitive use or Mr. or Sir. -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
