On Mon, 8 Jul 2002, Tom Lane wrote:
> Well, that's not unexpected; checkpoint is going to issue a deal of I/O
> and then sync() it. But that should *not* cause blockage of other
> backends; at worst they should slow down a bit due to I/O contention.
Well, depending on how the OS schedules writes, one process doing
a huge amount of writing might well slow down everything else a
lot, unless you've got a really good disk system.
But is it possible for a process to commit a transaction while a
checkpoint is in progress? That would mean that it's ok for the
checkpoint record to be after a bunch of transactions that are not
part of the checkpoint, right?
cjs
--
Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org
Don't you know, in this new Dark Age, we're all light. --XTC