On Fri, 2008-01-18 at 19:58 -0500, Tom Lane wrote: > In any case, 125 different zeroed pages is pretty hard to explain > by such a mechanism (especially if they were scattered rather than > in contiguous clumps). Can you show us the messages, so we can understand the distribution of the pages? Are there different numbers of rows in the two tables? Just a select count(*) might do, but any way you have of verifying data between the two systems would be very useful. The page numbers are identical between both systems, so use the contrib/pageinspect get_raw_page() function to record the contents on both systems before they diverge too much. (BTW, the final commit of those tools seems to have removed the docs I wrote for the original version and haven't been replaced with a README -- huh!?). create table bad_blocks as select <blockid> as blockid, get_raw_page(<relname>, <blockid>)::bytea as raw_page; I've got a bad feeling about rsync. Mason's recent problems used rsync and so far they are not properly explained, except as a hardware problem. -- Simon Riggs 2ndQuadrant http://www.2ndQuadrant.com ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster
