[pgpool-hackers: 4469] Re: abnormal behavior about PGPOOL RESET. and proposal a patch file.

[keiseo]

oh.. yes.. I agree. do the same thing which is already done is not desirable thing.

I just applied your patch and worked fine with your minimal code changes!

Initially, I used config_file_dir then realize it's unusable. but I couldn't think about change that initialization itself. So I use outer values which could be a side effect, that was not good I think.

anyway... thanks for the review!!



On Saturday, June 8th, 2024 at 17:25, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:

> Hello,
> 
> > Hello, and nice to meet you guys.
> > 
> > kashif_zeeshan(kashif.z at bitnine.net) found the abnormal query result with a specific query that is "PGPOOL RESET" query.
> > 
> > He shared this to me and I also had a same thing.
> > 
> > especially this bug only occurred when with watchdog is enabled.
> > 
> > I just do test some with regression test number 4 (watchdog) with fixing this one.
> > 
> > This is the text result.
> > 
> > before fixing.
> > -----------------
> > seo at seo-ThinkPad-E14-Gen-2:~$ psql -p 11000 test
> > psql (15.3)
> > Type "help" for help.
> > 
> > test=# PGPOOL RESET client_idle_limit;
> > ERROR: Pgpool node id file �0/pgpool_node_id does not exist
> > DETAIL: If watchdog is enable, pgpool_node_id file is required
> > test=# PGPOOL RESET client_idle_limit;
> > SET
> > ERROR: Pgpool node id file �y/pgpool_node_id does not exist
> > DETAIL: If watchdog is enable, pgpool_node_id file is required
> > message type 0x5a arrived from server while idle
> > message type 0x43 arrived from server while idle
> > message type 0x5a arrived from server while idle
> > test=#
> > 
> > after fixing.
> > -----------------
> > eo at seo-ThinkPad-E14-Gen-2:~$ psql -p 11000 test
> > psql (15.3)
> > Type "help" for help.
> > 
> > test=# PGPOOL RESET client_idle_limit;
> > SET
> > test=# PGPOOL RESET client_idle_limit;
> > SET
> > test=# PGPOOL RESET client_idle_limit;
> > SET
> > test=#
> > 
> > And also there's some explanation about why I do this and what I want to do in patch file.
> > 
> > Could you please review this patch?
> > 
> > If this is not a bug or only occurs for me, this patch should be abandoned. but that's also good for the pgpool because there's nothing matter on it!
> > 
> > oh well, I read the guideline but this is the first time sending a mail to hackers..
> > 
> > if I missed something, sorry about that and please tell me.
> 
> 
> Thank you for the problem report and the patch. I was able to
> reproduce the issue. Regarding your patch, I think there is no need
> to re-calculate the config file path in SetPgpoolNodeId(); it has been
> already calculated in ParseConfigFile(). All we need to do is, save
> the path in config_file_dir (and change config_file_dir from pointer
> to array).
> 
> Attached is the patch in this direction.
> 
> Best reagards,
> --
> Tatsuo Ishii
> SRA OSS LLC
> English: http://www.sraoss.co.jp/index_en/
> Japanese:http://www.sraoss.co.jp