Re: Questions about logicalrep_worker_launch()

On 2025/04/29 21:21, Amit Kapila wrote:
> On Mon, Apr 28, 2025 at 2:37 PM Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com> wrote:
>>
>> On 2025/04/26 3:03, Masahiko Sawada wrote:
>>> I agree with these changes.
>>>
>>> I think that while the changes for (2) should be for v19, the changes
>>> for (1) might be treated as a bug fix?
>>
>> Agreed. I've split the patch into two parts:
>>
>> 0001 is for (1) and is a bug fix that should be back-patched to v16,
>> where parallel apply workers were introduced. Since it didn't apply
>> cleanly to v16, I also created a separate patch specifically for v16.
>>
>
> The situation for parallel_apply workers is not as bad as for
> tablesync workers because even if the worker for parallel apply is not
> available, the apply worker will apply the changes by itself. OTOH, if
> the tablesync worker is not available, the tablesync will be pending
> till the time a worker for the same is not available. So, I am not
> sure if this is a clear cut bug which requires a fix in backbranches.

I'm fine with treating this as an improvement rather than a bug fix.
In any case, I've registered the patches for the next CommitFest.

The attached patches are the same as before, just rebased for the master branch.

>
> Additionally, shall we try to reproduce this case for parallel apply workers?

I noticed this issue while reading the code, so I haven't actually reproduced it.
Are you saying it's not possible to reproduce this in practice?

Regards,

--
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION