On 29.03.21 11:33, Amit Kapila wrote:
You don't need an additional callback for that if we do what I am
suggesting above.
Ah, are you suggesting a different change, then? To make two-phase
transactions always send PREPARE even if concurrently aborted? In that
case, sorry, I misunderstood.
I'm perfectly fine with that approach as well (even though it removes
flexibility compared to the concurrent abort callback, as the comment
above DecodePrepare indicates, i.e. "not impossible to optimize the
concurrent abort case").
One is you can try to test it, otherwise, there are comments atop
DecodePrepare() ("Note that we don't skip prepare even if have
detected concurrent abort because it is quite possible that ....")
which explains this.
Thanks for this pointer, very helpful.
Regards
Markus
