Обсуждение: disable wal

Is there a way to disable write ahead logging (WAL) in 7.1?

Thanks in advance.

Anuradha

Anuradha Ratnaweera wrote:
>
> Is there a way to disable write ahead logging (WAL) in 7.1?

Why do you want to do this?

Oliver

On Mon, 30 Apr 2001, Oliver Vecernik wrote:

> Anuradha Ratnaweera wrote:
> >
> > Is there a way to disable write ahead logging (WAL) in 7.1?
>
> Why do you want to do this?
>

Because I have a copy of my real server on a laptop and I do all the
development on it. I use pg_dumpall, diff and shell scripts to synchronize
the two databases, source code and mails.

Obviously, I don't want the data on the laptop to be reliable because even
if everything corrupts there, a fresh installation and a single run of the
sync scripts will take things back to normal. These scripts will _not_
corrupt the real server even if the data on the laptop is corrupt because
data is taken only _from_ the server and not the other way round.

WAL files seem to be a bottleneck for the laptop performance (not on the
real server). So I want to disable it on the laptop.

Regards,

Anuradha

There is no way to "disable WAL".  However, you might try running with
fsync turned off if you don't care about data integrity.

            regards, tom lane

Anuradha Ratnaweera wrote:
>
> On Mon, 30 Apr 2001, Oliver Vecernik wrote:
>
> > Anuradha Ratnaweera wrote:
> > >
> > > Is there a way to disable write ahead logging (WAL) in 7.1?
> >
> > Why do you want to do this?
> >
>
> Because I have a copy of my real server on a laptop and I do all the
> development on it. I use pg_dumpall, diff and shell scripts to synchronize
> the two databases, source code and mails.
>
> Obviously, I don't want the data on the laptop to be reliable because even
> if everything corrupts there, a fresh installation and a single run of the
> sync scripts will take things back to normal. These scripts will _not_
> corrupt the real server even if the data on the laptop is corrupt because
> data is taken only _from_ the server and not the other way round.
>
> WAL files seem to be a bottleneck for the laptop performance (not on the
> real server). So I want to disable it on the laptop.

If you don't need the other new features of 7.1 (outer joins, ...) you
can stick to 7.0.3.

Regards,
Oliver

I understand that that WAL is an integral part of 7.1 and "disabling" is
not possible. I also don't want to switch back to 7.0 becuse I _DO_ need
the new features.

If I don't care for reliability (for reasons I outlined earlier in this
thread), are there other parameters I can set to optimize for speed, such
as wal_sync_method? I already have turned off fsync and log_connections. I
can't turn on silent mode, because this machine is used for all testing
purposes and messages are useful.

Also, it is possible to avoid postgresql from trying to force the OS to
write data to the disk, so that WAL writes will be in disk cache for some
time?

Anuradha

On Mon, 30 Apr 2001, Tom Lane wrote:

> There is no way to "disable WAL".  However, you might try running with
> fsync turned off if you don't care about data integrity.
>
>             regards, tom lane
>

>
> I understand that that WAL is an integral part of 7.1 and "disabling" is
> not possible. I also don't want to switch back to 7.0 becuse I _DO_ need
> the new features.
>
> If I don't care for reliability (for reasons I outlined earlier in this
> thread), are there other parameters I can set to optimize for speed, such
> as wal_sync_method? I already have turned off fsync and log_connections. I
> can't turn on silent mode, because this machine is used for all testing
> purposes and messages are useful.
>
> Also, it is possible to avoid postgresql from trying to force the OS to
> write data to the disk, so that WAL writes will be in disk cache for some
> time?

Yes, as Tom said, you can use postgres -F flag to disable fsync writes,
even with WAL in 7.1.


>
> Anuradha
>
> On Mon, 30 Apr 2001, Tom Lane wrote:
>
> > There is no way to "disable WAL".  However, you might try running with
> > fsync turned off if you don't care about data integrity.
> >
> >             regards, tom lane
> >
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
>

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026