Re: effective_io_concurrency - Mailing list pgsql-admin

From Wells Oliver
Subject Re: effective_io_concurrency
Date
Msg-id CAOC+FBWjekHCW-njN2xKHDkfN_SF=RY3ZFWARhbjeqjBuNvZsA@mail.gmail.com
Whole thread Raw
In response to Re: effective_io_concurrency  (Ian Lawrence Barwick <barwick@gmail.com>)
List pgsql-admin
Looks like it's indeed 1 with an empty string in the RDS parameter groups. Thanks all.

On Wed, Jan 19, 2022 at 3:19 PM Ian Lawrence Barwick <barwick@gmail.com> wrote:
2022年1月20日(木) 3:24 Wells Oliver <wells.oliver@gmail.com>:
>
> Understood: wanted to double check that an outright empty value would be understood by the server as 0, or otherwise.

In standard PostgreSQL, if the configuration for an integer parameter
is empty or an
empty string, an error will be raised:

postgres=# ALTER SYSTEM SET effective_io_concurrency = '';
ERROR:  invalid value for parameter "effective_io_concurrency": ""

RDS is not standard PostgreSQL though, so something else might be
going on there.

Does:

  SELECT * FROM pg_settings WHERE name = 'effective_io_concurrency'

show anything?


Regards

Ian Barwick

> On Wed, Jan 19, 2022 at 9:46 AM Ron <ronljohnsonjr@gmail.com> wrote:
>>
>> On 1/19/22 11:40 AM, Wells Oliver wrote:
>>
>> Hi all-- I notice in the default PG 13 settings for Postgres on RDS, effective_io_concurrency is empty, and doing a "show effective_io_concurrency" also shows an empty value-- what's the effect of this? That PG assumes 0 for the setting?
>>
>>
>> https://www.postgresql.org/docs/13/runtime-config-resource.html
>>
>> "The allowed range is 1 to 1000, or zero to disable issuance of asynchronous I/O requests. Currently, this setting only affects bitmap heap scans."
>>
>> --
>> Wells Oliver
>> wells.oliver@gmail.com
>>
>>
>> --
>> Angular momentum makes the world go 'round.
>
>
>
> --
> Wells Oliver
> wells.oliver@gmail.com



--
EnterpriseDB: https://www.enterprisedb.com


--

pgsql-admin by date:

Previous
From: Ian Lawrence Barwick
Date:
Subject: Re: effective_io_concurrency
Next
From: Wells Oliver
Date:
Subject: COPY write load on primary impacting replica?