Hi Aidar, Thanks for testing both recovery paths and confirming the behavior.
> That said, I am not sure this is the right shape. The extra hunk fixes > a separate and pre-existing thing, and it probably wants its own thread and > a backpatch to 15 if it worths. So bundling it into this thread may be the > wrong > move. I did it only so the bot does not lose track of things. Nice catch on the disable_all_timeouts(false) issue. It appears to be independent of the standby timeout cleanup discussed in this thread and has different backpatching considerations. > If you or anyone else thinks the two should live apart, just say so and > I will split them: your v1 stays as is, and I will move the > disable_all_timeouts() fix to a new thread. +1 for splitting them. I suggest keeping Fujii-san’s standby timeout fix in this thread and moving the recovery-conflict issue to a separate thread and patch. That should allow the two changes to be reviewed, tested, and considered for backpatching independently. I have not reviewed the recovery-conflict patch in detail yet, but I would be happy to look at it separately once it is posted in its own thread. Best Regards, Nitin Jadhav Microsoft
