Good old corporate politics here. So some of this is just fishing to see to if anyone knows of anything that has changed from 8.1.7.4 to 9.2.0.3 with online index creation ... OR ... has there been a change made in the application (unbeknownst to us) that will crash into an online index creation. We're pretty sure the app doesn't try to do anything like grab a table lock ( ... we think ...), but it will certainly try to grab some "reasonable" ( ... we think ... ) row locks. There was no particular wait that was obvious during all the hullabaloo.
When this kind of thing is going on, you try to consult your Ouija board, examine the tea leaves, and shake the Magic 8 Ball in an attempt to get some gut feeling about what is going on without knowing exactly what the application is trying to do. The gut feeling I got was that whatever the app was doing to grab some rows was failing -- not just running slow or waiting, but failing. So either the app would loop around and try it again, or the app users were just banging away on some kind of retry button compounding the problem, and all of this grew into a big nasty snotball until the index creation was either killed, or got past its info gathering stage, after which things started working again. So there is some effort here to determine if there has been some change in the way 9.2.0.3 does online index creation, or can we say absolutely, positively it's exactly like 8.1.7.4; thus the problem must be a collision between the app and the index build; and it always would have been a collision both now and in the past. That way, we can focus on what the collision is and see if we can deal with it. > -----Original Message----- > > > I see, you can't find ot the waits because you don't want to ruin the > performance again on purpose. That makes sense! > > Online index rebuild has been problematic WRT the journal application > phase (among other problems such as abandoned journal segments after > cancel or failure) since it was introduced. On highly transactional > systems, a variety of waits, mostly library cache related, can occur > at the end of the index rebuild. Many more such problems are > addressed (fixed) in 9.2.0.3 than were in 8.1.7. > -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services --------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
