Re: PostgreSQL clustering with DRBD

From: Peter Eisentraut <peter_e(at)gmx(dot)net>
To: Serge Fonville <serge(dot)fonville(at)gmail(dot)com>
Cc: pgsql-general <pgsql-general(at)postgresql(dot)org>
Subject: Re: PostgreSQL clustering with DRBD
Date: 2009-02-11 11:26:22
Message-ID: 4992B5DE.6070301@gmx.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Serge Fonville wrote:
> Can PostgreSQL use DRBD as its storage?

Yes, many deployments use this.

> Since the in-memory database would be synchronized with the on-disk
> database.
> If this would be done with every query, this would greatly impact
> performance.

Performance with DRBD is usually acceptable, but you should test this
yourself in your setup.

> Since the cluster will be multi-master/dual-primary, do I need to have a
> separate block device for each PostgreSQL instance or can it use the
> DRBD device?

I don't understand how you want to have a multimaster PostgreSQL setup
with DRBD. DRBD can only be used for an inactive standby.

> I read mostly about MySQL clustering with DRBD and there the query cache
> should be disabled to make sure data is in-sync.
> To me it seems something similar would apply to PostgreSQL.

no

> I believe cybercluster is the most active and complete PostgreSQL
> clustering solution.
> My endgoal is a two node cluster with load sharing and fail over where
> both nodes can perform reads and writes.

That's not going to happen very easily, if at all.

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Serge Fonville 2009-02-11 11:52:10 Re: PostgreSQL clustering with DRBD
Previous Message Leif B. Kristensen 2009-02-11 10:35:54 Re: Difference between Windows pgsql and Linux pgsql?