[pgpool-hackers: 2566] Re: Watchdog fix commit to 3.4?

[Tatsuo Ishii]

Hi Usama,

Thank you for the confirmation!
I will go with this fix for 3.4.

>         Req_info->primary_node_id = -2;

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

> Hi Ishii San
> 
> On Sun, Oct 8, 2017 at 4:36 AM, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> 
>> Usama,
>>
>> You did not back patch following commit to 3.4:
>>
> 
> Watchdog was revamped in Pgpool-II 3.5 which modernised the watchdog
> communication protocol along with other things, And the older watchdogs
> lacked the necessary infrastructure to support this feature.
> So I only added this enhancement to Pgpool-II 3.5 and above
> 
> Thanks
> Best Regards
> Muhammad Usama
> 
> commit bda946e718fe6f3605eb7e82ada8754bd84a279c
>> Author: Muhammad Usama <m.usama at gmail.com>
>> Date:   Tue Aug 2 17:31:22 2016 +0500
>>
>>     Fix for 218: Inconsistent status of Postgresql nodes in pgPool
>> instances
>>     after restart. Watchdog not syncing status.
>>
>> In this commit you added:
>>
>>         /* initialize Req_info */
>>         Req_info->master_node_id = get_next_master_node();
>>         Req_info->conn_counter = 0;
>>         Req_info->switching = false;
>>         Req_info->request_queue_head = Req_info->request_queue_tail = -1;
>>         Req_info->primary_node_id = -2; <--- added
>>
>> So this was not added to 3.4.
>>
>> After that I committed following back to 3.4.
>> https://git.postgresql.org/gitweb/?p=pgpool2.git;a=commit;h=
>> e4ce880bd36b8f249bf693c086a1313148f3449a
>>
>> commit e4ce880bd36b8f249bf693c086a1313148f3449a
>> Author: Tatsuo Ishii <ishii at postgresql.org>
>> Date:   Wed May 10 08:30:17 2017 +0900
>>
>>     Fix corner case bug in Pgpool-II starting up.
>>
>>     It is possible that a failover request is accepted before primary node
>>     is searched.  This leads Pgpool-II to a strange state: there's no
>>     primary node if the failed node was a primary node (even if new
>>     primary node exists as a result of promotion of existing standby).
>>
>>     See [pgpool-hackers: 2321] for more details.
>>
>> In the commit I did this:
>>
>> +       /*
>> +        * if the primary node id is not loaded by watchdog, search for it
>> +        */
>> +       if (Req_info->primary_node_id < 0)
>> +       {
>> +               /* Save primary node id */
>> +               Req_info->primary_node_id = find_primary_node_repeatedly()
>> ;
>> +       }
>> +
>>
>> Since in 3.4 Req_info->primary_node_id is initialized to 0,
>> primary_node_repeatedly() is not called at the start up and this makes
>> 3.4 is not workable when the primary node is other than 0.
>>
>> I think we need to add following into 3.4:
>>
>>         Req_info->primary_node_id = -2;
>>
>> But I am not sure I am right because there maybe a reason why you
>> did't backpatch bda946e718fe6f3605eb7e82ada8754bd84a279c to 3.4.
>>
>> Can you please explain why you didn't back patch it?
>>
>> Best regards,
>> --
>> Tatsuo Ishii
>> SRA OSS, Inc. Japan
>> English: http://www.sraoss.co.jp/index_en.php
>> Japanese:http://www.sraoss.co.jp
>> _______________________________________________
>> pgpool-hackers mailing list
>> pgpool-hackers at pgpool.net
>> http://www.pgpool.net/mailman/listinfo/pgpool-hackers
>>