On Fri, Apr 14, 2017 at 9:19 PM, Petr Jelinek
<petr.jelinek@2ndquadrant.com> wrote:
> On 14/04/17 12:57, Masahiko Sawada wrote:
>> Hi,
>>
>> I noticed that the logical replication launcher uses
>> wal_retrieve_retry_interval as a interval of launching logical
>> replication worker process. This behavior is not documented and I
>> guess this is no longer consistent with what its name means.
>>
>
> Yes that was done based on reviews (and based on general attitude of not
> adding more knobs that are similar in meaning). It is briefly documented
> in the replication config section. Same is true for wal_receiver_timeout
> btw.
Thank you for the comment!
These two parameters are classed as a standby server parameter in the
document. We might want to fix it.
>> I think that we should either introduce a new GUC parameter (say
>> logical_replication_retry_interval?) for this or update the
>> description of wal_retrieve_retry_interval. IMO the former is better.
>>
>
> I am not quite sure adding more GUCs is all that great option. When
> writing the patches I was wondering if we should perhaps rename the
> wal_receiver_timeout and wal_retrieve_retry_interval to something that
> makes more sense for both physical and logical replication though.
>
It would work but IMO having multiple different behaviors for one GUC
parameter is not good design.
Regards,
--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center