[pgpool-hackers: 1691] Re: Question about new PGPOOL SET command

Tatsuo Ishii ishii at postgresql.org
Wed Jul 13 08:01:59 JST 2016


> On Thu, Jul 7, 2016 at 1:48 PM, Tatsuo Ishii <ishii at postgresql.org> wrote:
> 
>> Usama,
>>
>> In your implementation of new PGPOOL SET command in 3.6, config change
>> context is defined as enum:
>>
>> typedef enum
>> {
>>         CFGCXT_BOOT,
>>         CFGCXT_INIT,
>>         CFGCXT_RELOAD,
>>         CFGCXT_PCP,
>>         CFGCXT_SESSION
>> } ConfigContext;
>>
>> and each configuration parameter is assinged to one of them. Does the
>> order of enum value has any meaning ever? I mean, if certain variable
>> context is assiged to CFGCXT_RELOAD, then the variable can be changed
>> in CFGCXT_BOOT and CFGCXT_INIT context, for example.
>>
>> Yes, The order of enum value does matter, and the configuration variable
> with a specific config context can only be set/changed from the same or
> context with lower values.
> For example if the variable has the CFGCXT_RELOAD context, then it can only
> be changed using context <= CFGCXT_RELOAD (i.e. CFGCXT_BOOT, CFGCXT_INIT
> and CFGCXT_RELOAD) but not from the context having a value greater than
> variable context.

I see. Thanks.

> Do, you found any problem with it?

No. I was just thinking about adding new SET parameter in order to
change load balance node on the fly. That will be useful in the case
of planned switch over because with pgpool-II 3.6, session
disconnection can be avoided if the load balance node is not the
switch over target node.

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


More information about the pgpool-hackers mailing list