Simon Riggs <simon@2ndQuadrant.com> writes:
> These traces look weird. Look at the way the xid changes value as we
> move from call to call. It looks like something is screwy there. If
> those values are correct we should have failed an earlier assertion.
No, that's normal behavior on this platform + optimization setting.
Some of those registers have gotten re-used for other values. If
I were desperate to figure out how it got from point A to point B
I'd recompile with -O0, but this particular call stack doesn't seem
to hold any surprises: as you say, it seems to be trying to commit
an aborted xact. I looked far enough to see that the subxact ID
was a couple counts higher than the main, so I doubt that bad data
in the WAL record is the issue.
Are you able to reproduce the crash?
regards, tom lane