Hi Vijay, thanks for the tests, it's very helpful.
Just that, isn't it too extreme when one of the processes having problems accessing a snap file, it causes all other processes to be terminated? After all, most of the other processes do not need to access the snap files. Is there a way to fine tune this behaviour?
Thank you.
Regards,
Mike Yeap
I guess the best people to ask this question is the core team, I am just over a year in pg, I am not in a position to call this an issue with postgres unless I understand the underlying design and possible data corruption if any if it does not abort etc.
I'll leave it to the core team, then to give incorrect workaround.