Re: BUG #4879: bgwriter fails to fsync the file in recovery mode

Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
> Tom Lane wrote:
>> On the other point: are we going to eliminate mdunlink's isRedo
>> parameter? Maybe the better thing is to have its callers pass the value
>> of InArchiveRecovery?

> I think my initial analysis of this bug was bogus. There's nothing wrong
> with mdunlink() as it is, it's calling ForgetRelationFsyncRequests()
> before it unlinks anything, regardless of the isRedo parameter. In
> Fujii-san's scenario, it was just going to the pendingOpsTable of the
> startup process and not sent to bgwriter as it should. Setting
> pendingOpsTable=NULL when bgwriter is launched will fix that.

+1, I think that's okay. So to summarize the state of play, it seems
we have these issues:

* need to delete startup process's local pendingOpsTable once bgwriter
is launched, so that requests go to bgwriter instead

* need to push end-of-recovery checkpoint into bgwriter

* need to do something about IsRecovery tests that will now be executed
in bgwriter

* need to fix mistaken code in FinishPreparedTransaction

Anything else?

Are you (Heikki and Simon) in a position to finish these things today?
I know it's starting to get late on your side of the pond. Core have
already been discussing whether we have to abort the release for this.

regards, tom lane