[pgpool-general: 620] Re: ask some questions for pgpool-II

[Tatsuo Ishii]

I think you could do like this:

1) Use pgpool's native replication mode(not streaming replication)

2) Directly connect to master and create trigger which deletes rows
   being older than 3 month
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp

> sorry.
> source node is master
> target  node is  slave.
> master database sync to slave database, i  want to sync full data to
> slave,but need to delete some data on master,i write to some function
> including some delete command, i use black-function-list of replication
> mode that can filter this function on slave node.so some tables of master
> has recently three months data,some tables of slave node has all data .
> if i combine pgpool-II + PostgreSQL's streaming replication(async rep),both
> of master and slave data is equal. of course it is aysnc.
> 
> 2012/6/12 Tatsuo Ishii <ishii at postgresql.org>
> 
>> >> > I agree with you.i think that i find pgpool-II tool can filter some
>> >> > function (on source node delete data,on target node don't execute this
>> >> > command).
>>
>> I don't understand what you actually want to do. For example, what is
>> "source node", what is "target node"? Why do you need these exactly?
>> --
>> Tatsuo Ishii
>> SRA OSS, Inc. Japan
>> English: http://www.sraoss.co.jp/index_en.php
>> Japanese: http://www.sraoss.co.jp
>>
>> > combine pgpool-II + PostgreSQL's streaming replication(async rep). it is
>> > sync rep,but it can't filter some function ?
>> >
>> >
>> > 2012/6/12 Tatsuo Ishii <ishii at postgresql.org>
>> >
>> >> >> 3.slow node must be realy response client,if two nodes execute
>> commands
>> >> >> ,one fast,another slow,client gets time of result is from slow node.
>> >> >> the above is my worried solution!
>> >> >
>> >> > That's a trade off. If you prefer performance, you should use async
>> >> > replication. If you don't like delay of replication you should use
>> >> > sync replication. It's up to you.
>> >> > I agree with you.i think that i find pgpool-II tool can filter some
>> >> > function (on source node delete data,on target node don't execute this
>> >> > command).
>> >> > my requirement is source node that database keep  three months data
>> >> > recently at particular table,target node keep all data.
>> >> >  i find many tools,i think pgpoo-II is proper for my case,but it is
>> sync
>> >> > replication .
>> >> > if it will be async replication ,it is better!
>> >>
>> >> Actually you can. You can combine pgpool-II + PostgreSQL's streaming
>> >> replication(async rep).
>> >> --
>> >> Tatsuo Ishii
>> >> SRA OSS, Inc. Japan
>> >> English: http://www.sraoss.co.jp/index_en.php
>> >> Japanese: http://www.sraoss.co.jp
>> >>
>> >>
>> >> > thank you !
>> >> >
>> >> > 2012/6/12 Tatsuo Ishii <ishii at postgresql.org>
>> >> >
>> >> >> (Fowarded to mailing list to share info)
>> >> >>
>> >> >> > hello:
>> >> >> >      maybe i too worry it?
>> >> >> >  1.two replication node machine performance maybe not the same.so
>> one
>> >> >> node
>> >> >> > fast,another node slow,this soultion ,client will be feel it slow.
>> >> >>
>> >> >> No way.
>> >> >>
>> >> >> >  2.two replication node network maybe not the same,one network
>> >> fail,then
>> >> >> > the two node not sync,if it happen,how to resolve this solution ?
>> >> >> > i know pgpool has online recovery function,but how to monitor
>> status
>> >> of
>> >> >> the
>> >> >> > two node of pgpool?
>> >> >>
>> >> >> You can check pgpool log. Or you can email to admin in case of node
>> >> >> faiure by setting pgpool's failover script.
>> >> >>
>> >> >> > 3.slow node must be realy response client,if two nodes execute
>> >> commands
>> >> >> > ,one fast,another slow,client gets time of result is from slow
>> node.
>> >> >> > the above is my worried solution!
>> >> >>
>> >> >> That's a trade off. If you prefer performance, you should use async
>> >> >> replication.  If you don't like delay of replication you should use
>> >> >> sync replication. It's up to you.
>> >> >>
>> >> >> > please give me some suggestion how to tune pgpool-II replication ?
>> >> >>
>> >> >> You can adjust "weight" paramter to prevent the slow node receives
>> too
>> >> >> many queries.
>> >> >>
>> >> >> > please give me some atttion to use pgpool-II.
>> >> >> > how to monitor pgpool status using auto shell?
>> >> >>
>> >> >> Sorry, I don't know "auto shell".
>> >> >>
>> >> >> > i want to know that there are some case using pgpool-II
>> replication ?
>> >> >>
>> >> >> Check out pgpool-general mailing list archive. Recently a use case
>> has
>> >> >> been posted.
>> >> >> --
>> >> >> Tatsuo Ishii
>> >> >> SRA OSS, Inc. Japan
>> >> >> English: http://www.sraoss.co.jp/index_en.php
>> >> >> Japanese: http://www.sraoss.co.jp
>> >> >>
>> >> >> > tnank you very much ?
>> >> >> >
>> >> >> > 2012/6/11 Tatsuo Ishii <ishii at sraoss.co.jp>
>> >> >> >
>> >> >> >> > Hello Tatsuo:
>> >> >> >> >      my name is leo,my friend  Peng Sun introduces you to me.
>> >> >> >> >      i test pgpool-II replication function ,i find it is sync
>> >> >> relication
>> >> >> >> > from primary to other database,becase it is sync,i am worry it
>> that
>> >> >> >> stability
>> >> >> >> > and reliable.
>> >> >> >> > please give me some suggestions! Thank you very much!
>> >> >> >>
>> >> >> >> Why do you think sync replication is not reliable? Please
>> elaborate
>> >> >> >> your concerns so that I could answer your question.
>> >> >> >> --
>> >> >> >> Tatsuo Ishii
>> >> >> >> SRA OSS, Inc. Japan
>> >> >> >> English: http://www.sraoss.co.jp/index_en.php
>> >> >> >> Japanese: http://www.sraoss.co.jp
>> >> >> >>
>> >> >>
>> >>
>>