Horton

Thread

Commits

  1. Remove unused wait events.

  2. Add support for streaming to built-in logical replication.

  1. LogicalChanges* and LogicalSubxact* wait events are never reported

    Masahiro Ikeda <ikedamsh@oss.nttdata.com> — 2021-10-20T05:12:20Z

    Hi,

When I read the documents and source code of wait evens,
I found that the following wait events are never reported.

* LogicalChangesRead: Waiting for a read from a logical changes file.
* LogicalChangesWrite: Waiting for a write to a logical changes file.
* LogicalSubxactRead: Waiting for a read from a logical subxact file.
* LogicalSubxactWrite: Waiting for a write to a logical subxact file.


The wait events are introduced in the following patch.

 Add support for streaming to built-in logical replication.
 Amit Kapila on 2020/9/3 11:24:07
 464824323e57dc4b397e8b05854d779908b55304

I read the above discussion and found the wait events were reported at first.
But they seemed to be removed because they are not necessary because
BufFileWrite/BufFileRead are enough([1]).


If my understanding is right, it's better to remove them since they make
users confused. Please see the attached patch. I confirmed that to make
check-world passes all tests.

[1]
https://www.postgresql.org/message-id/CAA4eK1JV37jXUT5LeWzkBDNNnSntwQbLUZAj6m82QMiR1ZuuHQ%40mail.gmail.com

Regards,
-- 
Masahiro Ikeda
NTT DATA CORPORATION

  2. Re: LogicalChanges* and LogicalSubxact* wait events are never reported

    Michael Paquier <michael@paquier.xyz> — 2021-10-20T05:20:21Z

    On Wed, Oct 20, 2021 at 02:12:20PM +0900, Masahiro Ikeda wrote:
> If my understanding is right, it's better to remove them since they make
> users confused. Please see the attached patch. I confirmed that to make
> check-world passes all tests.

Yeah, I don't see the point in keeping these events around if they are
not used.  Perhaps Amit has some plans for them, though.
--
Michael

  3. Re: LogicalChanges* and LogicalSubxact* wait events are never reported

    Amit Kapila <amit.kapila16@gmail.com> — 2021-10-20T09:17:45Z

    On Wed, Oct 20, 2021 at 10:50 AM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Wed, Oct 20, 2021 at 02:12:20PM +0900, Masahiro Ikeda wrote:
> > If my understanding is right, it's better to remove them since they make
> > users confused. Please see the attached patch. I confirmed that to make
> > check-world passes all tests.
>
> Yeah, I don't see the point in keeping these events around if they are
> not used.  Perhaps Amit has some plans for them, though.
>

No, there is no plan for these. As far as I remember, during
development, we have decided to use BufFile interface and forgot to
remove these events which were required by the previous versions of
the patch. I'll take care of this.

Thanks for the report and patch!

-- 
With Regards,
Amit Kapila.

  4. Re: LogicalChanges* and LogicalSubxact* wait events are never reported

    Masahiro Ikeda <ikedamsh@oss.nttdata.com> — 2021-10-20T10:16:20Z

    
On 2021/10/20 18:17, Amit Kapila wrote:
> On Wed, Oct 20, 2021 at 10:50 AM Michael Paquier <michael@paquier.xyz> wrote:
>>
>> On Wed, Oct 20, 2021 at 02:12:20PM +0900, Masahiro Ikeda wrote:
>>> If my understanding is right, it's better to remove them since they make
>>> users confused. Please see the attached patch. I confirmed that to make
>>> check-world passes all tests.
>>
>> Yeah, I don't see the point in keeping these events around if they are
>> not used.  Perhaps Amit has some plans for them, though.
>>
> 
> No, there is no plan for these. As far as I remember, during
> development, we have decided to use BufFile interface and forgot to
> remove these events which were required by the previous versions of
> the patch. I'll take care of this.
> 
> Thanks for the report and patch!
Thanks for your replies and for handling it!

Regards,
-- 
Masahiro Ikeda
NTT DATA CORPORATION

  5. Re: LogicalChanges* and LogicalSubxact* wait events are never reported

    Amit Kapila <amit.kapila16@gmail.com> — 2021-10-21T08:40:55Z

    On Wed, Oct 20, 2021 at 3:46 PM Masahiro Ikeda <ikedamsh@oss.nttdata.com> wrote:
>
> On 2021/10/20 18:17, Amit Kapila wrote:
> > On Wed, Oct 20, 2021 at 10:50 AM Michael Paquier <michael@paquier.xyz> wrote:
> >>
> >> On Wed, Oct 20, 2021 at 02:12:20PM +0900, Masahiro Ikeda wrote:
> >>> If my understanding is right, it's better to remove them since they make
> >>> users confused. Please see the attached patch. I confirmed that to make
> >>> check-world passes all tests.
> >>
> >> Yeah, I don't see the point in keeping these events around if they are
> >> not used.  Perhaps Amit has some plans for them, though.
> >>
> >
> > No, there is no plan for these. As far as I remember, during
> > development, we have decided to use BufFile interface and forgot to
> > remove these events which were required by the previous versions of
> > the patch. I'll take care of this.
> >
> > Thanks for the report and patch!
> Thanks for your replies and for handling it!
>

Pushed!

-- 
With Regards,
Amit Kapila.

  6. Re: LogicalChanges* and LogicalSubxact* wait events are never reported

    Masahiro Ikeda <ikedamsh@oss.nttdata.com> — 2021-10-22T00:38:57Z

    
On 2021/10/21 17:40, Amit Kapila wrote:
> On Wed, Oct 20, 2021 at 3:46 PM Masahiro Ikeda <ikedamsh@oss.nttdata.com> wrote:
>>
>> On 2021/10/20 18:17, Amit Kapila wrote:
>>> On Wed, Oct 20, 2021 at 10:50 AM Michael Paquier <michael@paquier.xyz> wrote:
>>>>
>>>> On Wed, Oct 20, 2021 at 02:12:20PM +0900, Masahiro Ikeda wrote:
>>>>> If my understanding is right, it's better to remove them since they make
>>>>> users confused. Please see the attached patch. I confirmed that to make
>>>>> check-world passes all tests.
>>>>
>>>> Yeah, I don't see the point in keeping these events around if they are
>>>> not used.  Perhaps Amit has some plans for them, though.
>>>>
>>>
>>> No, there is no plan for these. As far as I remember, during
>>> development, we have decided to use BufFile interface and forgot to
>>> remove these events which were required by the previous versions of
>>> the patch. I'll take care of this.
>>>
>>> Thanks for the report and patch!
>> Thanks for your replies and for handling it!
>>
> 
> Pushed!

Thanks!

Regards,
-- 
Masahiro Ikeda
NTT DATA CORPORATION