Re: 100% failover + replication solution

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Brad Nicholson <bnichols(at)ca(dot)afilias(dot)info>
Cc: Moiz Kothari <moizpostgres(at)gmail(dot)com>, pgsql-admin(at)postgresql(dot)org
Subject: Re: 100% failover + replication solution
Date: 2006-11-18 00:30:53
Message-ID: 200611180030.kAI0UrS15276@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Brad Nicholson wrote:
> Excellent addition. I suggest in 24.1 - Shared Disk Failover something
> about the risk of having two postmasters come up on the same data
> directory, which is entirely possible if the old master doesn't release
> the disk in time.

I have added a mention of this.

---------------------------------------------------------------------------

>
>
>
> On Tue, 2006-11-14 at 15:11 -0500, Bruce Momjian wrote:
> > I wrote this for the 8.2 documentation. Let me know if it helps or you
> > have other suggestions:
> >
> > http://developer.postgresql.org/pgdocs/postgres/failover.html
> >
> > ---------------------------------------------------------------------------
> >
> > Moiz Kothari wrote:
> > > Guys,
> > >
> > > I have been thinking about this and wanted to see if it can be achived. I
> > > wanted to make a 100% failover solution for my postgres databases. The first
> > > thing that comes to my mind is doing it using WAL logs. Am attaching the
> > > diagram for which i will write more here.
> > >
> > > I was thinking if i can have Hotsync the databases using the WAL logs. If
> > > you see the architecture, the select goes to the SLAVE nodes and the INSERTS
> > > goes to the MASTER database, that means there should be no transactions
> > > happening on SLAVES (pure select). not even using "plpgsql". In this
> > > archecture i wont even mind restarting the SLAVE nodes if i need to.
> > >
> > > I was reading on PIT recovery on postgres, the only current problem i am
> > > facing with that is, i need to backup the database and move the slaves and
> > > the apply WAL logs to it. That could be huge amount of transfer and hence a
> > > longer downtime then moving 16mb WAL log files.
> > >
> > > Can someone help me out here, what i want is a continuous applying of WAL
> > > logs once i have brought the database up. I want to elimiate the backup of
> > > master to be restored to the slaves everytime i want to apply WAL to slaves.
> > > It would be real help if someone can tell me if it is possible to apply WAL
> > > logs to the slaves continously as soon as they are created on MASTER. I
> > > checked other tools available and thougth this would be best approach if it
> > > works.
> > >
> > > Awaiting reply soon.
> > >
> > > Regards,
> > > Moiz Kothari
> >
> > [ Attachment, skipping... ]
> >
> > >
> > > ---------------------------(end of broadcast)---------------------------
> > > TIP 5: don't forget to increase your free space map settings
> >
> --
> Brad Nicholson 416-673-4106
> Database Administrator, Afilias Canada Corp.
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings

--
Bruce Momjian bruce(at)momjian(dot)us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message rakesh kumar 2006-11-18 09:15:05 startup cost estimate
Previous Message Paul Forgey 2006-11-18 00:22:58 Re: remote connections to Windows based server