Andrew Dunstan <andrew@dunslane.net> writes:
> Well, the original code was put in for a reason, presumably that we were
> getting some stale data and wanted to exclude it. So I'm unwilling to
> throw it out altogether. If someone can propose a reasonable sanity
> check then I'm prepared to implement it.
Would it be practical to make the timeout configurable on a per-animal
basis? A small number of hours is certainly sufficient for most of the
critters, but for ones that are particularly slow or are doing CLOBBER
type tests, we could raise the limit.
regards, tom lane