Re: Compressing WAL

From: Qingqing Zhou
Subject: Re: Compressing WAL
Date: ,
Msg-id: d355d3$2ht4$
(view: Whole thread, Raw)
In response to: Compressing WAL  ("Jim C. Nasby")
List: pgsql-performance

Tree view

Compressing WAL  ("Jim C. Nasby", )
 Re: Compressing WAL  (Bruce Momjian, )
  Re: Compressing WAL  ("Jim C. Nasby", )
  Re: Compressing WAL  (Simon Riggs, )
   Re: Compressing WAL  (Bruce Momjian, )
 Re: Compressing WAL  ("Qingqing Zhou", )

""Jim C. Nasby"" <> writes
> Has anyone looked at compressing WAL's before writing to disk? On a
> system generating a lot of WAL it seems there might be some gains to be
> had WAL data could be compressed before going to disk, since today's
> machines are generally more I/O bound than CPU bound. And unlike the
> base tables, you generally don't need to read the WAL, so you don't
> really need to worry about not being able to quickly scan through the
> data without decompressing it.
> --

The problem is where you put the compression code? If you put it inside
XLogInsert lock or XLogWrite lock, which will hold the lock too long? Or
anywhere else?


pgsql-performance by date:

From: Tom Lane
Subject: Re: Never ending delete story
From: "Tambet Matiisen"
Subject: Re: performance - triggers, row existence etc.