Thread: describe-config issue

describe-config issue

From
vignesh C
Date:
Hi,

Postgres's describe-config option prints reset_val for int & real
configuration parameters which is not useful as it is not updated.
Printing boot_val is better in this case. reset_val is updated with
boot_val while the server is getting started but in case of postgres
--describe-config this value is not updated. I felt printing boot_val
is more appropriate in this case. Attached patch for the same.
Thoughts?

Regards,
Vignesh
EnterpriseDB: http://www.enterprisedb.com

Attachment

Re: describe-config issue

From
Tom Lane
Date:
vignesh C <vignesh21@gmail.com> writes:
> Postgres's describe-config option prints reset_val for int & real
> configuration parameters which is not useful as it is not updated.

Uh, what?

> Printing boot_val is better in this case.

Please defend that claim.  Otherwise this seems like a pretty
random change.

            regards, tom lane



Re: describe-config issue

From
vignesh C
Date:
On Wed, Sep 2, 2020 at 8:06 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:

> Please defend that claim.  Otherwise this seems like a pretty
> random change.

I had seen that there is discrepancy in postgres --describe-config & the value displayed from pg_settings like in the below case:
postgres=# select name,min_val, max_val, boot_val,reset_val from pg_settings where name = 'checkpoint_timeout';
        name        | min_val | max_val | boot_val | reset_val
--------------------+---------+---------+----------+-----------
 checkpoint_timeout | 30      | 86400   | 300      | 300
(1 row)

[vignesh@localhost bin]$ ./postgres --describe-config | grep checkpoint_timeout
checkpoint_timeout sighup Write-Ahead Log / Checkpoints INTEGER 0 30 86400 Sets the maximum time between automatic WAL checkpoints.

In the case of pg_settings we display 300 for boot_val/reset_val whereas in the case of describe-config we display 0, shouldn't it be 300 here?
Thoughts?

 Regards,
Vignesh
EnterpriseDB: http://www.enterprisedb.com