Mr. Riggs, Thank you for teaching me the following. I seem to have misunderstood. I'll learn more.
From: "Simon Riggs" <[EMAIL PROTECTED]> > On Fri, 2006-12-08 at 11:05 +0900, Takayuki Tsunakawa wrote: >> I understand that checkpoints occur during crash >> recovery and PITR, so time for those operations would get longer. > > A restorepoint happens during recovery, not a checkpoint. The recovery > is merely repeating the work of the checkpoint that occurred in the > original WAL stream. Elongating the checkpoint would not have any effect > on a restorepoint: we only record the checkpoint when it is complete and > we only create a restorepoint when we see the checkpoint record. Regards, ----- Original Message ----- From: "Simon Riggs" <[EMAIL PROTECTED]> To: "Takayuki Tsunakawa" <[EMAIL PROTECTED]> Cc: "ITAGAKI Takahiro" <[EMAIL PROTECTED]>; <pgsql-hackers@postgresql.org> Sent: Monday, December 11, 2006 6:30 PM Subject: Re: [HACKERS] Load distributed checkpoint > On Fri, 2006-12-08 at 11:05 +0900, Takayuki Tsunakawa wrote: >> I understand that checkpoints occur during crash >> recovery and PITR, so time for those operations would get longer. > > A restorepoint happens during recovery, not a checkpoint. The recovery > is merely repeating the work of the checkpoint that occurred in the > original WAL stream. Elongating the checkpoint would not have any effect > on a restorepoint: we only record the checkpoint when it is complete and > we only create a restorepoint when we see the checkpoint record. > > Crash recovery and PITR use almost exactly the same code path (by > design), so there isn't anything special to say about PITR either. > > -- > Simon Riggs > EnterpriseDB http://www.enterprisedb.com > > > ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
