Re: POC: Cleaning up orphaned files using undo logs

Dmitry Dolgov <9erthalion6@gmail.com> wrote:

> Yep, makes sense, thanks. I have few more questions:
>
> * The use case with orphaned files is working somewhat differently after
>   the rebase on the latest master, do you observe it as well? The
>   difference is ApplyPendingUndo -> SyncPostCheckpoint doesn't clean up
>   an orphaned relation file immediately (only later on checkpoint)
>   because of empty pendingUnlinks. I haven't investigated more yet, but
>   seems like after this commit:
>
>     commit 7ff23c6d277d1d90478a51f0dd81414d343f3850
>     Author: Thomas Munro <tmunro@postgresql.org>
>     Date:   Mon Aug 2 17:32:20 2021 +1200
>
>         Run checkpointer and bgwriter in crash recovery.
>
>         Start up the checkpointer and bgwriter during crash recovery (except in
>         --single mode), as we do for replication.  This wasn't done back in
>         commit cdd46c76 out of caution.  Now it seems like a better idea to make
>         the environment as similar as possible in both cases.  There may also be
>         some performance advantages.
>
>   something has to be updated (pendingOps are empty right now, so no
>   unlink request is remembered).

I haven't been debugging that part recently, but yes, this commit is relevant,
thanks for pointing that out! Attached is a patch that should fix it. I'll
include it in the next version of the patch series, unless you tell me that
something is still wrong.