Justin King <kingpin867@gmail.com> writes:
> I assume it would be related to the following:
> LOG: incorrect resource manager data checksum in record at 2D6/C259AB90
> since the walreceiver terminates just after this - but I'm unclear
> what precisely this means.
What it indicates is corrupt data in the WAL stream. When reading WAL
after crash recovery, we assume that that indicates end of WAL. When
pulling live data from a source server, it suggests some actual problem
... but killing the walreceiver and trying to re-fetch the data might
be a reasonable response to that. I'm not sure offhand what the startup
code thinks it's doing in this context. It might either be attempting
to retry, or concluding that it's come to the end of WAL and it ought
to promote to being a live server. If you don't see the walreceiver
auto-restarting then I'd suspect that the latter is happening.
regards, tom lane