Re: Replication

From: Hannu Krosing <hannu(at)skype(dot)net>
To: Markus Schiltknecht <markus(at)bluegap(dot)ch>
Cc: Fujii Masao <fujii(dot)masao(at)oss(dot)ntt(dot)co(dot)jp>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Replication
Date: 2006-08-23 11:20:06
Message-ID: 1156332007.2961.25.camel@localhost.localdomain
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Ühel kenal päeval, K, 2006-08-23 kell 13:09, kirjutas Markus
Schiltknecht:
> Hannu Krosing wrote:
> > But any sync _replication_ system will have severe impact on
> > performance. My guess is that for a full sync replication, going from 1
> > server to 2 will actually lower performance andsome small gains would be
> > possible only starting from 3rd server.
>
> Only testing will show concrete results, but for sure such a general
> answer does not make much sense. It all depends very much on the type of
> your load. Especially the read/write ratio is very important.

True.

But if you have very few writes, then there seems no reason to do sync
anyway.

> Another important factor is the amount of conflicting transactions.

That too, but just the need to do *any* locking on all nodes will
significantly slow down sync replication

> > My quess is based on using standard 100Gb ethernet. It may be possible
> > to do better on some more advanced interconnects like myrinet.
>
> 100Gb ethernet already sounds very advanced... ;-)

Yeah, I meant some mix of 100Mb and 1Gb ;)

> Regards
>
> Markus
>
--
----------------
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia

Skype me: callto:hkrosing
Get Skype for free: http://www.skype.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Markus Schiltknecht 2006-08-23 11:36:37 Re: Replication
Previous Message Markus Schiltknecht 2006-08-23 11:09:36 Re: Replication