Greg Stark <stark@mit.edu> writes:
> So I think I've come up with a scenario that could cause this. I don't
> think it's exactly what happened here but maybe something analogous
> happened with our base backup restore.
I agree it seems like a good idea for XLogReadBufferExtended to defend
itself against successive P_NEW calls possibly not returning consecutive
pages. However, unless you had an operating-system-level crash on the
master, this isn't sufficient to explain the problem. We'd need also a
plausible theory about how a base backup could've left us with short
segments in a large relation.
> (Or maybe the hot backup
> process could just catch the files in this state if a table is rapidly
> growing and it doesn't take care to avoid picking up new files that
> appear after it starts?)
That's a possible explanation I guess, but it doesn't seem terribly
probable from a timing standpoint. Also, you should be able to gauge
the probability of this theory from knowledge of the application ---
are the bloated relations all ones that would've been growing *very*
rapidly during the base backup?
regards, tom lane