> Question is why there seems to be such a difference in size between > the master and the slave. ... > cls.listings rows:1608406 pages:335434 shouldbe:222939 (1.5X) wasted ... > cls.listings rows:1599038 pages:292044 shouldbe:222460 (1.3X) wasted ... > Yes I checked the autovacs and they ran 2 hours apart but during a > time of low traffic, this would not explain the pretty significant > difference in wasted space.. I'm wondering where I should start to > look and or if I should waste any effort on tracking this down?
Try ANALYZING both right before running the check, to get the best row and pages estimates. At the very least, the number of rows should match much closer than they are. It may just be that the vacuum on the slave side did not run as recently as the master, or could not completely remove all dead tuples. Compare ANALYZE VERBOSE on both sides for a closer look at the differences. -- Greg Sabino Mullane [EMAIL PROTECTED] End Point Corporation
signature.asc
Description: PGP signature
_______________________________________________ Slony1-general mailing list [email protected] http://lists.slony.info/mailman/listinfo/slony1-general
