[pgpool-general: 5971] Re: pgpool load balancing problem

[Tatsuo Ishii]

Not sure what you mean. It's a source code and you can copy into
whatever directory. Probably safest option is, copy into different
directory and compile all over again.

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

> I downloaded it, but should I copy those files to the original pool
> directory(the current pool directory that has the bug) or I should replace
> it and compile and all over again ?
> 
> 
> 2018-03-04 15:32 GMT+02:00 Mariel Cherkassky <mariel.cherkassky at gmail.com>:
> 
>> Is the link you specified in the end of your comment is the 3.7.2 version
>> without the bug ?
>>
>> 2018-03-02 1:10 GMT+02:00 Tatsuo Ishii <ishii at sraoss.co.jp>:
>>
>>> So you are surely bitten by the bug in 3.7.2.
>>>
>>> In other message you failed to incomporate the patch. Maybe executing
>>> "make clean" before compiling helps you.
>>>
>>> If you are not familiar with patching, you could obtain whole source
>>> code from:
>>> https://git.postgresql.org/gitweb/?p=pgpool2.git;a=shortlog;
>>> h=refs/heads/V3_7_STABLE
>>>
>>> In this window the top most row is the latest entry for 3.7.x. You hit
>>> "snapshot" link to get the source code tar ball (.tgz).
>>>
>>> Current direct link to the tar ball is:
>>> https://git.postgresql.org/gitweb/?p=pgpool2.git;a=snapshot;
>>> h=8593af680d9c575c23a823489ecfa693749811f3;sf=tgz
>>>
>>> Another workaround is, swapping the entry for node 0 and node 1 in
>>> your pgpool.conf.
>>>
>>> Best regards,
>>> --
>>> Tatsuo Ishii
>>> SRA OSS, Inc. Japan
>>> English: http://www.sraoss.co.jp/index_en.php
>>> Japanese:http://www.sraoss.co.jp
>>>
>>> > pool_version
>>> >
>>> > ----------------------
>>> >
>>> > 3.7.2 (amefuriboshi)
>>> >
>>> > (1 row)
>>> >
>>> >
>>> >
>>> >
>>> >
>>> > psql -d postgres -U postgres -h 127.0.0.1  -p 9999 -c "show pool_nodes";
>>> >
>>> >
>>> >
>>> > node_id |    hostname     | port | status | lb_weight |  role   |
>>> > select_cnt | load_balance_node | replication_delay
>>> >
>>> > ---------+-----------------+------+--------+-----------+----
>>> -----+------------+-------------------+-------------------
>>> >
>>> > 0       | pg1 | 5432 | up     | 0.333333  | standby | 22565      |
>>> > true              | 0
>>> >
>>> > 1       | pg2 | 5432 | up     | 0.333333  | primary | 10416      |
>>> > false             | 0
>>> >
>>> > 2       | pg3 | 5432 | up     | 0.333333  | standby | 7686       |
>>> > false             | 0
>>> >
>>> > (3 rows)
>>> >
>>> >
>>> > 2018-03-01 6:23 GMT+02:00 Tatsuo Ishii <ishii at sraoss.co.jp>:
>>> >
>>> >> > I found a problem that had a big impact on our servers. I'm using
>>> pgpool
>>> >> as
>>> >> > a load balancer and pool connections manager. When a user tries to
>>> run a
>>> >> > select query it redirects the query to the slave. If the user tries
>>> after
>>> >> > that to run dml or ddl command he get error :
>>> >> >
>>> >> > ERROR: cannot execute CREATE TABLE in a read-only transaction
>>> >> > SQL state: 25006
>>> >> >
>>> >> > Is there any way to solve this issue ? It means that once the pool
>>> >> > redirected your session to the slaves you cant run any ddl/dml
>>> actions ..
>>> >>
>>> >> Normaly Pgpool-II redirects any ddl/dml query to primary.  Can you
>>> >> show me following query results? I would like to confirm your problem.
>>> >>
>>> >> show pool_version;
>>> >> show pool_nodes;
>>> >>
>>> >> Best regards,
>>> >> --
>>> >> Tatsuo Ishii
>>> >> SRA OSS, Inc. Japan
>>> >> English: http://www.sraoss.co.jp/index_en.php
>>> >> Japanese:http://www.sraoss.co.jp
>>> >>
>>>
>>
>>