FW: Performance problems on Informix 7.x on HP UX 10.0x on HP9000

Use create temp table ... with no log. -----Original Message----- From: harryjohnson@my-deja.com [mailto:harryjohnson@my-deja.com] Sent: Tuesday, March 07, 2000 1:44 PM To: informix-list@iiug.org Subject: Re: Performance problems on Informix 7.x on HP UX 10.0x on HP9000 If you cannot take logging off your database, you might think about adding more logs to your system. Whatever you do, DO NOT CHANGE the long transaction high water mark!!!!!!!! I have had several clients do this. They had to get the GURU's at Informix to fix their "frozen" system. Harry Johnson Keene Information Systems, Inc. In article <JJJw4.7539$6b1.120741@news1.online.no>, "Frank T. Ramb'l" <news@d-fect.com> wrote: > Hi. > > At work we have a large database which includes the tables tracked_ids > and image_ids among others. > > tracked_ids have the following rows: > scanner integer > image_id char(33) > chk_count integer > scandate date > > image_ids consist of: > image_id char(33) > > I run the following script: > === > begin work; > set lock mode to wait;> > create table ftr_temp_strings (id char(11));> > insert into ftr_temp_strings select image_id[16,26] from image_ids;> > update tracked_ids set chk_count=-1 > where image_id in (select id from ftr_temp_strings);> > delete from image_ids > where image_id[16,26] in (select image_id from tracked_ids where> chk_count=-1); > > delete from tracked_ids where chk_count=-1;> > drop table ftr_temp_strings;> > commit work; > === > > The problem is that it stops with errormessage "458: Long transaction > aborted." in the "insert into ftr_temp_strings"-line. > > I have also tried using this version: > === > begin work; > set lock mode to wait;> > update tracked_ids set chk_count=-1 > where image_id in (select image_id[16,26] from image_ids);> > delete from image_ids > where image_id[16,26] in (select image_id from tracked_ids where> chk_count=-1); > > delete from tracked_ids where chk_count=-1;> > commit work; > === > > The problem here is the same. It times out on the first > update-statement. That is why we made the other version because that is > somewhat faster. > > Any help, ideas, hints whatever that can help me with this problem is > very much apreciated. > > If there is any more information you want or need to answer me, just ask > and I will try to find out. > > Many thanks in advance, > Frank > > Sent via Deja.com http://www.deja.com/ Before you buy.