Path: news.netfront.net!news.glorb.com!postnews.google.com!u10g2000prn.googlegroups.com!not-for-mail From: gazzag Newsgroups: comp.databases.oracle.server Subject: Re: application problems Date: Mon, 11 Feb 2008 09:25:43 -0800 (PST) Organization: http://groups.google.com Lines: 41 Message-ID: References: <022efb99-9bc3-4974-ba2d-28396ef2847e@s8g2000prg.googlegroups.com> NNTP-Posting-Host: 193.195.187.198 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Trace: posting.google.com 1202750743 13985 127.0.0.1 (11 Feb 2008 17:25:43 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Mon, 11 Feb 2008 17:25:43 +0000 (UTC) Complaints-To: groups-abuse@google.com Injection-Info: u10g2000prn.googlegroups.com; posting-host=193.195.187.198; posting-account=XNB8mAgAAACJCSJ3htUgwEp6l3Q9aDMr User-Agent: G2/1.0 X-HTTP-UserAgent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; Meridio for Excel 4.4.727; Meridio for PowerPoint 4.4.727; Meridio for Word 4.4.727; .NET CLR 1.1.4322; .NET CLR 2.0.50727),gzip(gfe),gzip(gfe) Xref: news.netfront.net comp.databases.oracle.server:182334 On 11 Feb, 16:51, Ben wrote: > 10.2.0.2 Ent Ed AIX 5.2 64 bit > > We have a vendor app that we recently upgraded and in turn moved it > from SQLServer to an Oracle database environment. All was well for > about a month then we started seeing random application failures. I've > looked at everything I know to look at in the database and don't see > any kind of indication that it is something within the database > causing the issues. In my opinion the app wasn't written well for an > Oracle db to begin with, they don't use any bind variables. It is a > fairly small (600M) oltp type of application but does have quite a few > transactions. The SGA is about 600M total and I haven't seen any > errors in the alert log to make me think that we are hitting any kind > of memory errors related to SQL cursors using up all the sga, or core > dumps, or anything for that matter to make me think there should be an > issue. The parse percentages are ridiculously bad, but even with that > I don't see extreme waits related to parse times. > > Wouldn't I be seeing alert log errors if something in the sga was a > miss and causing client side problems? > > Their client side logging is pretty lack luster as well, just cryptic > ole errors. In my experience, most database applications are pretty poorly written. This is largely down to the software house's desire to create "database unaware" programs. The rationale here being that it doesn't matter whether a potential customer wants to use Oracle, DB2, SQL Server or even mySQL, for that matter, as their backend, the applicaton itself won't need a rewrite; it'll behave just as poorly across the board. Anyhow, you say "random application failures". I'm afraid that this doesn't give us too much to go on. What do you mean when you say "failure"? HTH -g