Mark Nickel <mnickel@dunsirn.com> writes:
> Would having the --enable-debug in the build of postgres have helped to
> show more information in the stack trace? Would that extra information
> been more helpful to determining the problem in this case?
Yes, and maybe. If you're using gcc then there's really no reason
not to --enable-debug. The executable files will balloon by several
megabytes worth of debugging information, but this is just extra disk
space; there's no run-time penalty to having it there, and someday you
might be glad you did. (On a non-gcc compiler there may be some runtime
penalty, since --enable-debug defeats code optimization in most other
compilers.)
> I'm off to plan an upgrade to 7.1.3 :)
Good idea. I can think of several scenarios that can lead to corrupted
index in 7.0.*, but we've tried to defend against each of them in 7.1.
regards, tom lane
