For Andrew Cilia - Select & logical logs - mail bounces

FYI. } } This is a MIME-encapsulated message } } --RAA17084.838740596/link.net.mt } } The original message was received at Tue, 30 Jul 1996 17:26:01 +0200 (MET) } from hp.com [15.255.152.4] } } ----- The following addresses have delivery notifications ----- } <acilia@mail.ptl.com.mt> (unrecoverable error) } } ----- Transcript of session follows ----- } 553 mail.link.net.mt. config error: mail loops back to me (MX problem?) } 554 <acilia@mail.ptl.com.mt>... Local configuration error } 553 MAILER-DAEMON... Invalid recipient...User does not exist } } --RAA17084.838740596/link.net.mt } Content-Type: message/delivery-status } } Reporting-MTA: dns; link.net.mt } Received-From-MTA: DNS; hp.com } Arrival-Date: Tue, 30 Jul 1996 17:26:01 +0200 (MET) } } Final-Recipient: RFC822; acilia@mail.ptl.com.mt } Action: failed } Status: 5.5.0 } Remote-MTA: DNS; mail.link.net.mt } Last-Attempt-Date: Tue, 30 Jul 1996 17:29:56 +0200 (MET) } } --RAA17084.838740596/link.net.mt } } I am currently investigating an incident on a 5.03 ONLINE } engine where a select statement with a correlated subquery } seems to populate the logical logs. Reputable sources tell } me that this is standard procedure on complicated queries. } Can anyone confirm this? Where is it documented? } Precisely under what conditions are logs used by queries. } Why does it need to be this way? } } Customer is more than disgruntled since the log requirements } of this infamous query far exceeds those of all other processes } in the system!!! } EVEN if you don't have logging on. Any statement which alters a schema results in a logical log entry. Temporary tables are included in this lot. Unless working off of a clustered index (and potentially even then) any ORDER BY (for example) will require a temporary table. Subqueries are especially bad since you may generate a temporary table for each potentially returned row from the main select. I could be out to lunch on this last one, but that sure seems to be the way it works. (I'm sure I'll be corrected if I am wrong). The workaround in a non-logging environment? Turn on continuous logging to /dev/null. cheers j. _____________________________________________________________________________ Jack Parker - Hewlett Packard, DMD/IS Boise, Idaho, USA jparker@boi.hp.com Currently on loan to PLD/PE _____________________________________________________________________________ If anything can go wrong, fix it. To hell with Murphy. _____________________________________________________________________________ Any opinions expressed herein are my own and not those of my employers. _____________________________________________________________________________