| From: | Koichi Suzuki <suzuki(dot)koichi(at)oss(dot)ntt(dot)co(dot)jp> |
|---|---|
| To: | Martijn van Oosterhout <kleptog(at)svana(dot)org> |
| Cc: | Jim Nasby <decibel(at)decibel(dot)org>, PGSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Archive log compression keeping physical log available in the crash recovery |
| Date: | 2007-02-13 02:02:14 |
| Message-ID: | 45D11C26.6040200@oss.ntt.co.jp |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Further, because pgbench writes many zero values to fixed length
columns, gzip can achieve better compression. There're another
suggestion to test with longer checkpoint interval. I will post the
result.
Thanks.
Martijn van Oosterhout wrote:
> On Fri, Feb 09, 2007 at 01:00:10PM +0900, Koichi Suzuki wrote:
>> Further, we can apply gzip to this archive (2.36GB). Final size is
>> 0.75GB, less than one sixtieth of the original WAL.
>
> Note that if you were compressing on the fly, you'll have to tell gzip
> to regularly flush its buffers to make sure all the data actually hits
> disk. That cuts into your compression ratio...
>
> Have a nice day,
--
Koichi Suzuki
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2007-02-13 02:08:05 | Re: [PATCHES] Bad bug in fopen() wrapper code |
| Previous Message | Tom Lane | 2007-02-13 01:55:12 | Re: Foreign keys for non-default datatypes, redux |