RE: any impact to DML sql statements when checkpoint starting

["James Pang (chaolpan)" <chaolpan@xxxxxxxxx>]

When Postgres checkpoint,  it always write buffers from pg shared_buffers  to  OS filesystem cache,  after that,  it finally make a system call  fsync , that will flush dirty blocks from OS filesystem cache to disk, right?

 

Thanks,

 

James

 

From: David G. Johnston <david.g.johnston@xxxxxxxxx>
Sent: Friday, February 4, 2022 8:50 AM
To: James Pang (chaolpan) <chaolpan@xxxxxxxxx>
Cc: pgsql-admin@xxxxxxxxxxxxxxxxxxxx
Subject: Re: any impact to DML sql statements when checkpoint starting

 

On Thu, Feb 3, 2022 at 5:45 PM James Pang (chaolpan) <chaolpan@xxxxxxxxx> wrote:

So the main concern about checkpoint impact  is IO,  in order to reduce the IO by checkpoint, in addition to postgres checkpoint parameters. Linux(Redhat 8), some OS kernel parameters need to be set ,right?

 

vm.dirty_expire_centisecs=500

vm.dirty_writeback_centisecs=250

vm.dirty_ratio=30

vm.dirty_background_ratio = 10

 

 

I understand the big picture (more-or-less), not tuning specifics.  But for a checkpoint we don't really give the O/S much control - it needs to do fsync when we ask for it.  It's in between the explicit issuance of checkpoints that the algorithms written into PostgreSQL and the O/S come into play to hopefully spread out the load over more time and keep the active memory holding useful data.

 

David J.