Re: log_checkpoint's "WAL file(s) added" is misleading to the point of uselessness - Mailing list pgsql-hackers

From Bossart, Nathan
Subject Re: log_checkpoint's "WAL file(s) added" is misleading to the point of uselessness
Date
Msg-id BC5A4B01-537B-4FE5-A0B7-40D8DC9E0AA4@amazon.com
Whole thread Raw
In response to Re: log_checkpoint's "WAL file(s) added" is misleading to the point of uselessness  (Fujii Masao <masao.fujii@oss.nttdata.com>)
List pgsql-hackers
On 7/26/21, 5:23 PM, "Fujii Masao" <masao.fujii@oss.nttdata.com> wrote:
> On 2021/07/27 5:27, Bossart, Nathan wrote:
>> +1.  I was confused by this when working on a WAL pre-allocation
>> patch [0].  Perhaps it could be replaced by a new parameter and a new
>> field in pg_stat_wal.  How about something like log_wal_init_interval,
>> where the value is the minimum amount of time between reporting the
>> number of WAL segments created since the last report?
>
> You mean to introduce new GUC like log_wal_init_interval and that
> the number of WAL files created since the last report will be logged
> every that interval? But isn't it better and simpler to just expose
> the accumulated number of WAL files created, in pg_stat_wal view
> or elsewhere? If so, we can easily get to know the number of WAL files
> created in every interval by checking the view and calculating the diff.

I agree with you about adding a new field to pg_stat_wal.  The
parameter would just be a convenient way of logging this information
for future reference.  I don't feel strongly about the parameter if
you think the pg_stat_wal addition is enough.

Nathan


pgsql-hackers by date:

Previous
From: Justin Pryzby
Date:
Subject: Re: needless complexity in StartupXLOG
Next
From: Andres Freund
Date:
Subject: Re: log_checkpoint's "WAL file(s) added" is misleading to the point of uselessness