Path: newssvr20.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!newsfeed.cwix.com!news.maxwell.syr.edu!postnews1.google.com!not-for-mail From: pagesflames@usa.net (Dusan Bolek) Newsgroups: comp.databases.oracle.server Subject: Re: Oracle 9.2 Log Miner Scripts? Date: 11 Nov 2003 05:10:59 -0800 Organization: http://groups.google.com Lines: 52 Message-ID: <1e8276d6.0311110510.692e51cd@posting.google.com> References: <1e8276d6.0311070129.702f2fe3@posting.google.com> <1e8276d6.0311100040.a372cd4@posting.google.com> <3fb05402$0$9221$afc38c87@news.optusnet.com.au> NNTP-Posting-Host: 195.47.102.140 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1068556259 32535 127.0.0.1 (11 Nov 2003 13:10:59 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Tue, 11 Nov 2003 13:10:59 +0000 (UTC) Xref: newssvr20.news.prodigy.com comp.databases.oracle.server:247443 "Howard J. Rogers" wrote in message news:<3fb05402$0$9221$afc38c87@news.optusnet.com.au>... > "Dusan Bolek" wrote in message > news:1e8276d6.0311100040.a372cd4@posting.google.com... > > > It is not about cheapness. We're using EE everywhere, but this > > particular need could be only satisfied with LOGICAL stand-by, which > > is not very usable on 9.2. > > > > -- > > Dusan Bolek > > I'm intrigued, Dusan: can you elaborate as to *why* it's not very usable? > Problems you've encountered? Bugs? Restrictions in functionality? Weirdness > of your particular setup?? > > Genuine interest this end... you can mail me offline if you'd prefer. OK, I will tell you why I've decided no to use Logical standby. In the beginning when I performed initial analysis how to solve this business need and logical stand-by was definitely an option. However, from beginning I had some prejudice against logical stand-by. The reason was that Oracle 9i rel 2 was the first release where logical stand-by was implemented. From my past experiences I'm somehow reluctant to use in production brand new features from Oracle, because sometimes Oracle needs at least one other release to master it. So I've done some background search and found: a) there are several not very nice bugs associated with logical stand-by (some of them were patched later on) b) some people, who tried to use logical stand-by, reported that LS's maturity is not at desired level c) I've got a response from Oracle itself that using of logical stand-by in highly critical productional environment is possible, but not recommended. Another problem was that I wanted to have open possibility to decrease network load in case that we would realize that too much data is shipped through our lines. This open possibility was to use interconnects between disk arrays (EMC's Symmetrixes) to ship redo from productional to archive database. Data Guard supplies redo logs to standby using NET8, which is running above TCP/IP protocol (and several others), but cannot be (nicely) implemented above proprietary interconnect (not suprising). Summary: Maybe it was possible to use logical standby to do our job. However, I was not enough brave (or mad?) to risk this on productional system which is one the most critical to us. Maybe you can blame me for a lack of courage, but I still think that my decision was the right one. And is working too ... -- Dusan Bolek