Tom Lane escribió:
> Given this, and the index corruption you showed before (the wrong
> sibling link, which would represent index breakage quite independent of
> what was in the heap), and the curious contents of your WAL files
> (likewise not explainable by anything going wrong within a table),
> I'm starting to think that Occam's razor says you've got hardware
> problems. Or maybe a kernel-level bug that is causing writes to get
> discarded.
FWIW, on a customer's setup we've seen something that looks very much
like this: table pages being rewritten by something that I couldn't
identify, but which certainly weren't normal table or index pages. On
looking at your report and diagnosis, I'm inclined to think that they
were WAL records due to the similarity and regularity. (I'll take a
look again).
I also wrote them off as hardware failure or kernel bugs.
I don't think we've been seen the problem again. But what happened in
the end was that the customer asked us whether we could make something
about CRC'ing the files we use in order to detect system failure
proactively.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support