[pgpool-hackers: 3536] Re: Refactoring running mode in Pgpool-II

[Tatsuo Ishii]

>> > Yes, Its would be a nice feature and it would create more clarity in
>> > configuration. Few comments/suggests.
>> > 1.
>> > To me "running_mode" seems bit generic, and it might be confusing whether
>> > its is something related to how pgpool application was started ( like we
>> > have pgpool -m smart/fast/etc ),
>> > We might have same feature of 'pgpool start' as well (
>> > headless/light/integrated etc. )
>> >
>> > Instead of 'running_mode' we can have 'replication_setup' or
>> > 'replication_setup_mode'  with following value:
>> > streaming
>> > logical
>> > native
>> > slony
>> > raw
>>
>> Makes sense. I prefer 'replication_setup_mode'.
>>
> 
>  replication_setup_mode seems good, I am also thinking about something like
> "backend_clustering_mode" or "backend_clustering_setup"
> What do you think?

That's even better I think since not all mode does replication (hint:
raw mode). I am going to use 'backend_clustering_mode' and start
development of patches.

>> > 2.
>> > Just an idea, we should develop this in extensibility in mind. In
>> feature,
>> > if we have some new kind of replication mode, pgpool should be able to
>> > accommodate it with minimum changes ( or my be user can provide their own
>> > replication mode somehow, like we have memcached extensibility )
>>
>> That's an interesting idea. Maybe something like *preload_libraries in
>> PostgreSQL?
>>
> 
> That's an interesting idea but we should also keep a close eye on the
> performance impact of such a change

Right.

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp