ARCH vs LGWR for redo forwarding to standby, was: About standby redo logs

  • From: Carel-Jan Engel <cjpengel.dbalert@xxxxxxxxx>
  • To: greg.loughmiller@xxxxxxxxxxxx
  • Date: Wed, 05 Jan 2005 20:26:21 +0100

Using LGWR for redo forwarding has a big advantage: How many logswitches
do you have per hour? Two, three? How many transactions will there be on
average in each redo log file? That is the amount of transactions that
can be lost when the primary fails. And of course, Murphy will arrange
the failure to happen just before an ordinary logswitch. 

When LGWR forwards the redo, it will jeapordize just a couple of
transactions, or even zero, depending on the mode you chose. You can
have the LGWR forwarding redo in synchronous/affirmed mode, and that
will guarantee your transactions to be stored on the standby.
Asynchronous mode will result in just a few tenths of seconds, or just e
few seconds, of redo to be buffered, still reducing the amount of lost
data quite significantly. Most of the OLTP systems with Data Guard I set
up are using synchronous mode, because the systems and network don't
notice any performance loss. I think no data loss after take-over is
better than possible data-loss, whatever little that may be. Data loss
is leaving the users uncertain about which  transactions are lost,
taking quite some time to find out and manually recover.

HTH, Carel-Jan 

If you think education is expensive, try ignorance. (Derek Bok)

Upcoming appearances: 

      * Jan 27, 2005: London, UKOUG Unix SIG: Data Guard Best Practices 
      * Feb 9-10, 2005: Denver, RMOUG Training Days: Data Guard
        Performance Issues 
      * Mar 6-10, 2005: Dallas, Hotsos Symposium: Data Guard Performance

DBA!ert, Independent Oracle consultancy 
Kastanjelaan 61C
2743 BX  Waddinxveen
The Netherlands
tel. +31 (0) 182 64 04 28
fax +31 (0) 182 64 04 29
e-mail info.dbalert@xxxxxxxxx

> If I may ask a similar question, yet not exactly related to this
> thread..
> What are the advantages of using LGWR as compared to the ARCH process
> for
> the transport mechanism of the redo data to the standby database?
> Thanks
> Greg


Other related posts:

  • » ARCH vs LGWR for redo forwarding to standby, was: About standby redo logs