Vivek Khera wrote:
> And the winner is... checkpoint_segments.
> Restore of a significanly big database (~19.8GB restored) shows nearly
> no time difference depending on sort_mem when checkpoint_segments is
> large. There are quite a number of tables and indexes. The restore
> was done from a pg_dump -Fc dump of one database.
> All tests with 16KB page size, 30k shared buffers, sort_mem=8192, PG
> 7.4b2 on FreeBSD 4.8.
> 3 checkpoint_segments restore time: 14983 seconds
> 50 checkpoint_segments restore time: 11537 seconds
> 50 checkpoint_segments, sort_mem 131702 restore time: 11262 seconds
With the new warning about too-frequent checkpoints, people have actual
feedback to encourage them to increase checkpoint_segments. One issue
is that it is likely to recommend increasing checkpoint_segments during
restore, even if there is no value to it being large during normal
server operation. Should that be decumented?
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
In response to
pgsql-performance by date
|Next:||From: Vivek Khera||Date: 2003-09-23 14:27:55|
|Subject: Re: restore time: sort_mem vs. checkpoing_segments|
|Previous:||From: Nick Fankhauser||Date: 2003-09-22 20:42:27|
|Subject: How to make n_distinct more accurate. |