Re: pg_ctl -D?
От | Richard Broersma Jr |
---|---|
Тема | Re: pg_ctl -D? |
Дата | |
Msg-id | 20061004223804.52232.qmail@web31814.mail.mud.yahoo.com обсуждение исходный текст |
Ответ на | Re: pg_ctl -D? (Ray Stell <stellr@cns.vt.edu>) |
Ответы |
Re: pg_ctl -D?
|
Список | pgsql-novice |
> This may come back to bite me some other way. Having all the data on one > drive doesn't make any sense to me at all. Would this not cause i/o bottlenecks? Yes it probably would create i/o bottlenecks to some degree. This is why using a multi-disc hardware raid-array with a BBU is so popular. There is alot of discussion of this on the preformance list. A raid 10 array will allow for the increase I/O through-put. I have seen advice to the effect that hand partioning your database accross muliple drives many not produce desirable results and is a bad idea. It seems and it is hard to balance tuple growth evenly across all of the drives. Also most of your data writes are on one table, then you are limited to the i/o bandwidth of one drive anyway. However, a better/safer way to relieve i/o conjestion is to move your wal-archieve directory to another drive. Regards, Richard Broersma Jr.
В списке pgsql-novice по дате отправления: