Re: [HACKERS] A few notes
Mike Mascari [EMAIL PROTECTED] writes: ... is it possible the GEQO threshold, as a default, is too low? It's been on the TODO list for awhile to investigate whether the current default is still appropriate. A lot of planner details have changed since we last twiddled it, and you're not the first to suggest that it may need to be twiddled again. But no one's provided more than anecdotal evidence. If you want to do some measurements to determine where the best crossover point is now, go to it ... regards, tom lane ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
[HACKERS] A few notes
Hello. I thought I'd just toss out a few thoughts: 1) Should a link to the release changes for 7.3.3 be on the website? I had to look into the web-interface of CVS to see what was actually changed. 2) It would be nice if some regular performance tests could be done upon every release on some stock machine whose configuration never changes to give some numerical hints as to the value of an upgrade. 3) I got bit by using the explicit join syntax just like Thomas Lockhart had predicted. I then removed the syntax to let the planner do its job. Queries which took around 10 seconds took 5 minutes. I then disabled GEQO and the queries ran in around a second. I noticed that the explicit join syntax will no longer confine planning choices in 7.4, but is it possible the GEQO threshold, as a default, is too low? Mike Mascari [EMAIL PROTECTED] ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
