RE: Acceptable latency for LGWR-SYNC and LGWR-ASYNC
Date: Fri, 23 Mar 2012 10:53:30 -0400
I think Niall is heading toward what latency being worse than would cause you to question the set-up for a mis-configuration given certain infrastucture, while I *think* Hemant is fishing for the acceptable hang on commit confirmation delay to dictate the network infrastructure requirements.
If I'm correct, that delay in turn is variable depending on what the database is being used for. If we're talking about a potential hang in a web transaction confirmation I'm not sure whether that plays in the same space as recent measurements about web site conversions. Presumably if your site displays quickly and they've already invested in making choices to the point of committing a transaction the delay from that point forward is not as sensitive and you have a captive audience anyway.
I think it all depends on what you are making people wait for or whether the lwgr latency is pacing batch job transaction throughput and how close to acid you need your system to be in site disasters.
Regards, and apologies in advance if I have misconstrued your inquiry.
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Niall Litchfield
Sent: Friday, March 23, 2012 6:02 AM
Subject: Re: Acceptable latency for LGWR-SYNC and LGWR-ASYNC
I think you'd need to give a few more details, like network details, redo generation rate, distance between sites etc. On Fri, Mar 23, 2012 at 1:10 AM, Hemant K Chitale <hemantkchitale_at_gmail.com>wrote:
> What would be acceptable latency for DataGuard using LGWR instead of
> ARCH (in 10.2) ?
> Hemant K Chitale
> sent from my smartphone
-- Niall Litchfield Oracle DBA http://www.orawin.info -- http://www.freelists.org/webpage/oracle-l -- http://www.freelists.org/webpage/oracle-lReceived on Fri Mar 23 2012 - 09:53:30 CDT