postgres-xc-general Mailing List for Postgres-XC

Thanks Ashutosh for the correct answer.    So far, Postgres-XC allows constraints which can be enforced without additional infrastructure, I mean, without visiting other datanodes to check if the constraint is maintained.    I'm afraid Alex referred to a distributed table from replicated tables, which cannot be enforced with the current XC infrastructure.     Of course, the opposite should work, referring replicated tables from a distributed table.

Regards;
---
Koichi Suzuki

On 2013/10/29, at 15:29, Ashutosh Bapat <ash...@en...<mailto:ash...@en...>>
 wrote:

Hi Alex,
The error comes because, the referenced key is not guaranteed to on the same node where the row referencing it lies. Thus either of following should hold for adding a foreign key.
1. Both the tables (referenced and referencer) should be replicated on the same set of nodes
2. Both the tables should be distributed by in the same fashion on referenced and referencer columns resp.
3. The referencer table is distributed and referenced table is replicated on all the nodes where the referencer table is distributed.


Koichi Suzuki says that this is current XC limitation, I tried to Change the tables distribution to "replication". I can change the layout, but the problem is when I want to add the foreign key, I get the same error. How can I do to have the keys in my system?

The system has at least 10 nodes…



Regards,

Alex Hudson S.


------------------------------------------------------------------------------
Android is increasing in popularity, but the open development platform that
developers love is also attractive to malware creators. Download this white
paper to learn more about secure code signing practices that can help keep
Android apps secure.
http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk
_______________________________________________
Postgres-xc-general mailing list
Pos...@li...<mailto:Pos...@li...>
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general




--
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company
------------------------------------------------------------------------------
Android is increasing in popularity, but the open development platform that
developers love is also attractive to malware creators. Download this white
paper to learn more about secure code signing practices that can help keep
Android apps secure.
http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk_______________________________________________
Postgres-xc-general mailing list
Pos...@li...
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general

Hi Alex,
The error comes because, the referenced key is not guaranteed to on the
same node where the row referencing it lies. Thus either of following
should hold for adding a foreign key.
1. Both the tables (referenced and referencer) should be replicated on the
same set of nodes
2. Both the tables should be distributed by in the same fashion on
referenced and referencer columns resp.
3. The referencer table is distributed and referenced table is replicated
on all the nodes where the referencer table is distributed.

Koichi Suzuki says that this is current XC limitation, I tried to
Change the tables distribution to "replication". I can change the
layout, but the problem is when I want to add the foreign key, I get
the same error. How can I do to have the keys in my system?****
>
> The system has at least 10 nodes… ****
>
> ** **
>
> Regards,****
>
> ** **
>
> Alex Hudson S.****
>
> ** **
>
>
> ------------------------------------------------------------------------------
> Android is increasing in popularity, but the open development platform that
> developers love is also attractive to malware creators. Download this white
> paper to learn more about secure code signing practices that can help keep
> Android apps secure.
> http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

Hi guys,

 

I could really use your help. 
I'm a sysadmin. I'm not a DBA, and I don't have any experience with
PostgreSQL.
 
I'm trying PGXC as a Cluster solution for the PGSQL DB Component in our
Car's assistance infrastructure.
The problem I'm currently having, is that when I try to set a foreign key in
any table.
 
"ERROR:  Cannot create foreign key whose evaluation cannot be enforced to
remote nodes"
 
Koichi Suzuki says that this is current XC limitation, I tried to Change the
tables distribution to "replication". I can change the layout, but the
problem is when I want to add the foreign key, I get the same error. How can
I do to have the keys in my system?
The system has at least 10 nodes. 
 
Regards,

 

Alex Hudson S.

 

Ashutosh,

 First, just wanted to confirm about the streaming aspect. There are two
parts to this: one the streaming aggregate operators and making query
planning cognizant about it.
As I said, the whole query planner and how to extend it is not so clear.
You did  give some pointers about how to go about doing this. However,
someone with not complete knowledge of the planner subsystem would have
still have diffculty.

I can help with the documentation first about how to extend the planner if
you are interested and can give  pointers in codebase which needs to
touched in order to achieve this.

-Sandeep





On Thu, Oct 17, 2013 at 8:19 AM, Ashutosh Bapat <
ash...@en...> wrote:

> That sounds a good idea. Can you please provide a patch?
>
>
> On Thu, Oct 17, 2013 at 5:41 PM, Sandeep Gupta <gup...@gm...>wrote:
>
>>  I maybe missing something. Is it the case that postgres or pgxc doesn't
>> support
>> streaming aggregates? That would allow aggregation over sorted streams to
>> produce a
>> aggregated stream that is also sorted.
>> However, I am not too sure about this.
>>
>> -Sandeep
>>
>>
>>
>>
>> On 10/17/2013 07:13 AM, Ashutosh Bapat wrote:
>>
>> That's not possible right now. For grouping, either the grouped input
>> from the datanode gets shuffled at the coordinator or ordered on the
>> grouping column. Thus even if we get the ordered intput from datanode on a
>> particular column that order is disturbed at the coordinator because of
>> grouping.
>>
>>
>> On Thu, Oct 17, 2013 at 4:37 PM, Sandeep Gupta <gup...@gm...>wrote:
>>
>>>  How about a plan where the datanodes perform the sort as well and the
>>> coordinator performs a sorted merge?
>>>
>>> Are such plans not part of the query planner?
>>>
>>> -Sandeep
>>>
>>>
>>> On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
>>>
>>> There is GROUP BY clause that needs to be evaluated before the result
>>> can be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.
>>>
>>>
>>>  On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta <gup...@gm...
>>> > wrote:
>>>
>>>>  Hi Ashutosh,
>>>>
>>>>
>>>>    Attached below is the query and the corresponding query plan. I am
>>>> using version 1.1.
>>>>  Thanks for taking a look at this.
>>>>
>>>>  -Sandeep
>>>>
>>>> SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>                                      QUERY PLAN (Coordinator)
>>>>
>>>>  Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>>>>    Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>>>>    Sort Key: a.exposed_time
>>>>    ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>>>>          Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>>>>          ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>>>>                Output: a.exposed_time, (count(a.pid))
>>>>                Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>>>>                Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>>>> (9 rows)
>>>>
>>>>                                      QUERY PLAN (Datanode)
>>>>
>>>>  GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>>>>    Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>>>>    ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>>>>          Output: a.exposed_time, a.pid
>>>>          ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>>>>                Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>>>>                Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>>>>          ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>>>>                Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>>>>                Index Cond: (d.pid = a.pid)
>>>>                Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>>>> (11 rows)
>>>>
>>>>
>>>>
>>>>
>>>> On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
>>>> ash...@en...> wrote:
>>>>
>>>>> Sandeep,
>>>>> It would be nice if you mention the version of XC in your mail. Sort
>>>>> push down is available from 1.1 onwards. If you do not see sort getting
>>>>> pushed down in 1.1, please report detailed definitions of the tables, query
>>>>> and the EXPLAIN output.
>>>>>
>>>>>
>>>>>  On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <
>>>>> gup...@gm...> wrote:
>>>>>
>>>>>>    Hi,
>>>>>>
>>>>>>   In an another query that requires the result to be aggregated and
>>>>>> ordered by a field (lets say timeo)
>>>>>> the query planner currently pulls  the results and then performs a
>>>>>> sort with hash aggregate.
>>>>>>
>>>>>>  The table at the datanodes are clustered by timeo. I was wondering
>>>>>> if it possible
>>>>>> for query planner to push down the order by clause at the datanode
>>>>>> and then perform
>>>>>>  sort-merge aggregate at the coordinator. Surely, that would be a
>>>>>> better query plan.
>>>>>>
>>>>>>  We have tried enable_sort=off etc. but that doesn't work.
>>>>>>
>>>>>>  Thanks.
>>>>>>  Sandeep
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> October Webinars: Code for Performance
>>>>>> Free Intel webinars can help you accelerate application performance.
>>>>>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the
>>>>>> most from
>>>>>> the latest Intel processors and coprocessors. See abstracts and
>>>>>> register >
>>>>>>
>>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>>>>> _______________________________________________
>>>>>> Postgres-xc-general mailing list
>>>>>> Pos...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Best Wishes,
>>>>> Ashutosh Bapat
>>>>> EnterpriseDB Corporation
>>>>> The Postgres Database Company
>>>>>
>>>>
>>>>
>>>
>>>
>>> --
>>> Best Wishes,
>>> Ashutosh Bapat
>>> EnterpriseDB Corporation
>>> The Postgres Database Company
>>>
>>>
>>>
>>
>>
>> --
>> Best Wishes,
>> Ashutosh Bapat
>> EnterpriseDB Corporation
>> The Postgres Database Company
>>
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>

That sounds a good idea. Can you please provide a patch?


On Thu, Oct 17, 2013 at 5:41 PM, Sandeep Gupta <gup...@gm...>wrote:

>  I maybe missing something. Is it the case that postgres or pgxc doesn't
> support
> streaming aggregates? That would allow aggregation over sorted streams to
> produce a
> aggregated stream that is also sorted.
> However, I am not too sure about this.
>
> -Sandeep
>
>
>
>
> On 10/17/2013 07:13 AM, Ashutosh Bapat wrote:
>
> That's not possible right now. For grouping, either the grouped input from
> the datanode gets shuffled at the coordinator or ordered on the grouping
> column. Thus even if we get the ordered intput from datanode on a
> particular column that order is disturbed at the coordinator because of
> grouping.
>
>
> On Thu, Oct 17, 2013 at 4:37 PM, Sandeep Gupta <gup...@gm...>wrote:
>
>>  How about a plan where the datanodes perform the sort as well and the
>> coordinator performs a sorted merge?
>>
>> Are such plans not part of the query planner?
>>
>> -Sandeep
>>
>>
>> On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
>>
>> There is GROUP BY clause that needs to be evaluated before the result can
>> be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.
>>
>>
>>  On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta <gup...@gm...>wrote:
>>
>>>  Hi Ashutosh,
>>>
>>>
>>>    Attached below is the query and the corresponding query plan. I am
>>> using version 1.1.
>>>  Thanks for taking a look at this.
>>>
>>>  -Sandeep
>>>
>>> SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>>>
>>>
>>>
>>>
>>>
>>>                                      QUERY PLAN (Coordinator)
>>>
>>>  Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>>>    Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>>>    Sort Key: a.exposed_time
>>>    ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>>>          Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>>>          ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>>>                Output: a.exposed_time, (count(a.pid))
>>>                Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>>>                Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>>> (9 rows)
>>>
>>>                                      QUERY PLAN (Datanode)
>>>
>>>  GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>>>    Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>>>    ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>>>          Output: a.exposed_time, a.pid
>>>          ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>>>                Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>>>                Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>>>          ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>>>                Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>>>                Index Cond: (d.pid = a.pid)
>>>                Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>>> (11 rows)
>>>
>>>
>>>
>>>
>>> On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
>>> ash...@en...> wrote:
>>>
>>>> Sandeep,
>>>> It would be nice if you mention the version of XC in your mail. Sort
>>>> push down is available from 1.1 onwards. If you do not see sort getting
>>>> pushed down in 1.1, please report detailed definitions of the tables, query
>>>> and the EXPLAIN output.
>>>>
>>>>
>>>>  On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <
>>>> gup...@gm...> wrote:
>>>>
>>>>>    Hi,
>>>>>
>>>>>   In an another query that requires the result to be aggregated and
>>>>> ordered by a field (lets say timeo)
>>>>> the query planner currently pulls  the results and then performs a
>>>>> sort with hash aggregate.
>>>>>
>>>>>  The table at the datanodes are clustered by timeo. I was wondering if
>>>>> it possible
>>>>> for query planner to push down the order by clause at the datanode and
>>>>> then perform
>>>>>  sort-merge aggregate at the coordinator. Surely, that would be a
>>>>> better query plan.
>>>>>
>>>>>  We have tried enable_sort=off etc. but that doesn't work.
>>>>>
>>>>>  Thanks.
>>>>>  Sandeep
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> October Webinars: Code for Performance
>>>>> Free Intel webinars can help you accelerate application performance.
>>>>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the
>>>>> most from
>>>>> the latest Intel processors and coprocessors. See abstracts and
>>>>> register >
>>>>>
>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>>>> _______________________________________________
>>>>> Postgres-xc-general mailing list
>>>>> Pos...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Best Wishes,
>>>> Ashutosh Bapat
>>>> EnterpriseDB Corporation
>>>> The Postgres Database Company
>>>>
>>>
>>>
>>
>>
>> --
>> Best Wishes,
>> Ashutosh Bapat
>> EnterpriseDB Corporation
>> The Postgres Database Company
>>
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

I maybe missing something. Is it the case that postgres or pgxc doesn't 
support
streaming aggregates? That would allow aggregation over sorted streams 
to produce a
aggregated stream that is also sorted.
However, I am not too sure about this.

-Sandeep



On 10/17/2013 07:13 AM, Ashutosh Bapat wrote:
> That's not possible right now. For grouping, either the grouped input 
> from the datanode gets shuffled at the coordinator or ordered on the 
> grouping column. Thus even if we get the ordered intput from datanode 
> on a particular column that order is disturbed at the coordinator 
> because of grouping.
>
>
> On Thu, Oct 17, 2013 at 4:37 PM, Sandeep Gupta 
> <gup...@gm... <mailto:gup...@gm...>> wrote:
>
>     How about a plan where the datanodes perform the sort as well and the
>     coordinator performs a sorted merge?
>
>     Are such plans not part of the query planner?
>
>     -Sandeep
>
>
>     On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
>>     There is GROUP BY clause that needs to be evaluated before the
>>     result can be ordered. Thus GROUP BY is sent to the datanode but
>>     not ORDER BY.
>>
>>
>>     On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta
>>     <gup...@gm... <mailto:gup...@gm...>> wrote:
>>
>>         Hi Ashutosh,
>>
>>
>>           Attached below is the query and the corresponding query
>>         plan. I am using version 1.1.
>>         Thanks for taking a look at this.
>>
>>         -Sandeep
>>
>>         SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>>
>>
>>
>>
>>                                             QUERY PLAN (Coordinator)
>>
>>           Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>>             Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>>             Sort Key: a.exposed_time
>>             ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>>                   Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>>                   ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>>                         Output: a.exposed_time, (count(a.pid))
>>                         Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>>                         Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>>         (9 rows)
>>
>>                                               QUERY PLAN (Datanode)
>>
>>           GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>>             Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>>             ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>>                   Output: a.exposed_time, a.pid
>>                   ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>>                         Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>>                         Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>>                   ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>>                         Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>>                         Index Cond: (d.pid = a.pid)
>>                         Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>>         (11 rows)
>>
>>
>>
>>
>>         On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat
>>         <ash...@en...
>>         <mailto:ash...@en...>> wrote:
>>
>>             Sandeep,
>>             It would be nice if you mention the version of XC in your
>>             mail. Sort push down is available from 1.1 onwards. If
>>             you do not see sort getting pushed down in 1.1, please
>>             report detailed definitions of the tables, query and the
>>             EXPLAIN output.
>>
>>
>>             On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta
>>             <gup...@gm...
>>             <mailto:gup...@gm...>> wrote:
>>
>>                 Hi,
>>
>>                  In an another query that requires the result to be
>>                 aggregated and ordered by a field (lets say timeo)
>>                 the query planner currently pulls  the results and
>>                 then performs a sort with hash aggregate.
>>
>>                 The table at the datanodes are clustered by timeo. I
>>                 was wondering if it possible
>>                 for query planner to push down the order by clause at
>>                 the datanode and then perform
>>                 sort-merge aggregate at the coordinator. Surely, that
>>                 would be a better query plan.
>>
>>                 We have tried enable_sort=off etc. but that doesn't
>>                 work.
>>
>>                 Thanks.
>>                 Sandeep
>>
>>
>>
>>                 ------------------------------------------------------------------------------
>>                 October Webinars: Code for Performance
>>                 Free Intel webinars can help you accelerate
>>                 application performance.
>>                 Explore tips for MPI, OpenMP, advanced profiling, and
>>                 more. Get the most from
>>                 the latest Intel processors and coprocessors. See
>>                 abstracts and register >
>>                 http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>                 _______________________________________________
>>                 Postgres-xc-general mailing list
>>                 Pos...@li...
>>                 <mailto:Pos...@li...>
>>                 https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>
>>
>>
>>
>>             -- 
>>             Best Wishes,
>>             Ashutosh Bapat
>>             EnterpriseDB Corporation
>>             The Postgres Database Company
>>
>>
>>
>>
>>
>>     -- 
>>     Best Wishes,
>>     Ashutosh Bapat
>>     EnterpriseDB Corporation
>>     The Postgres Database Company
>
>
>
>
> -- 
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company

This description is for distributed tables. For replicated tables,
everything shippable is shipped.


On Thu, Oct 17, 2013 at 4:43 PM, Ashutosh Bapat <
ash...@en...> wrote:

> That's not possible right now. For grouping, either the grouped input from
> the datanode gets shuffled at the coordinator or ordered on the grouping
> column. Thus even if we get the ordered intput from datanode on a
> particular column that order is disturbed at the coordinator because of
> grouping.
>
>
> On Thu, Oct 17, 2013 at 4:37 PM, Sandeep Gupta <gup...@gm...>wrote:
>
>>  How about a plan where the datanodes perform the sort as well and the
>> coordinator performs a sorted merge?
>>
>> Are such plans not part of the query planner?
>>
>> -Sandeep
>>
>>
>> On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
>>
>> There is GROUP BY clause that needs to be evaluated before the result can
>> be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.
>>
>>
>>  On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta <gup...@gm...>wrote:
>>
>>>  Hi Ashutosh,
>>>
>>>
>>>    Attached below is the query and the corresponding query plan. I am
>>> using version 1.1.
>>>  Thanks for taking a look at this.
>>>
>>>  -Sandeep
>>>
>>> SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>>>
>>>
>>>
>>>
>>>
>>>                                      QUERY PLAN (Coordinator)
>>>
>>>  Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>>>    Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>>>    Sort Key: a.exposed_time
>>>    ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>>>          Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>>>          ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>>>                Output: a.exposed_time, (count(a.pid))
>>>                Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>>>                Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>>> (9 rows)
>>>
>>>                                      QUERY PLAN (Datanode)
>>>
>>>  GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>>>    Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>>>    ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>>>          Output: a.exposed_time, a.pid
>>>          ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>>>                Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>>>                Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>>>          ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>>>                Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>>>                Index Cond: (d.pid = a.pid)
>>>                Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>>> (11 rows)
>>>
>>>
>>>
>>>
>>> On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
>>> ash...@en...> wrote:
>>>
>>>> Sandeep,
>>>> It would be nice if you mention the version of XC in your mail. Sort
>>>> push down is available from 1.1 onwards. If you do not see sort getting
>>>> pushed down in 1.1, please report detailed definitions of the tables, query
>>>> and the EXPLAIN output.
>>>>
>>>>
>>>>  On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <
>>>> gup...@gm...> wrote:
>>>>
>>>>>    Hi,
>>>>>
>>>>>   In an another query that requires the result to be aggregated and
>>>>> ordered by a field (lets say timeo)
>>>>> the query planner currently pulls  the results and then performs a
>>>>> sort with hash aggregate.
>>>>>
>>>>>  The table at the datanodes are clustered by timeo. I was wondering if
>>>>> it possible
>>>>> for query planner to push down the order by clause at the datanode and
>>>>> then perform
>>>>>  sort-merge aggregate at the coordinator. Surely, that would be a
>>>>> better query plan.
>>>>>
>>>>>  We have tried enable_sort=off etc. but that doesn't work.
>>>>>
>>>>>  Thanks.
>>>>>  Sandeep
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> October Webinars: Code for Performance
>>>>> Free Intel webinars can help you accelerate application performance.
>>>>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the
>>>>> most from
>>>>> the latest Intel processors and coprocessors. See abstracts and
>>>>> register >
>>>>>
>>>>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>>>> _______________________________________________
>>>>> Postgres-xc-general mailing list
>>>>> Pos...@li...
>>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Best Wishes,
>>>> Ashutosh Bapat
>>>> EnterpriseDB Corporation
>>>> The Postgres Database Company
>>>>
>>>
>>>
>>
>>
>> --
>> Best Wishes,
>> Ashutosh Bapat
>> EnterpriseDB Corporation
>> The Postgres Database Company
>>
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>



-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

That's not possible right now. For grouping, either the grouped input from
the datanode gets shuffled at the coordinator or ordered on the grouping
column. Thus even if we get the ordered intput from datanode on a
particular column that order is disturbed at the coordinator because of
grouping.


On Thu, Oct 17, 2013 at 4:37 PM, Sandeep Gupta <gup...@gm...>wrote:

>  How about a plan where the datanodes perform the sort as well and the
> coordinator performs a sorted merge?
>
> Are such plans not part of the query planner?
>
> -Sandeep
>
>
> On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
>
> There is GROUP BY clause that needs to be evaluated before the result can
> be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.
>
>
>  On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta <gup...@gm...>wrote:
>
>>  Hi Ashutosh,
>>
>>
>>    Attached below is the query and the corresponding query plan. I am
>> using version 1.1.
>>  Thanks for taking a look at this.
>>
>>  -Sandeep
>>
>> SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>>
>>
>>
>>
>>
>>                                      QUERY PLAN (Coordinator)
>>
>>  Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>>    Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>>    Sort Key: a.exposed_time
>>    ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>>          Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>>          ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>>                Output: a.exposed_time, (count(a.pid))
>>                Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>>                Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>> (9 rows)
>>
>>                                      QUERY PLAN (Datanode)
>>
>>  GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>>    Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>>    ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>>          Output: a.exposed_time, a.pid
>>          ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>>                Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>>                Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>>          ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>>                Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>>                Index Cond: (d.pid = a.pid)
>>                Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>> (11 rows)
>>
>>
>>
>>
>> On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
>> ash...@en...> wrote:
>>
>>> Sandeep,
>>> It would be nice if you mention the version of XC in your mail. Sort
>>> push down is available from 1.1 onwards. If you do not see sort getting
>>> pushed down in 1.1, please report detailed definitions of the tables, query
>>> and the EXPLAIN output.
>>>
>>>
>>>  On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <gup...@gm...
>>> > wrote:
>>>
>>>>    Hi,
>>>>
>>>>   In an another query that requires the result to be aggregated and
>>>> ordered by a field (lets say timeo)
>>>> the query planner currently pulls  the results and then performs a sort
>>>> with hash aggregate.
>>>>
>>>>  The table at the datanodes are clustered by timeo. I was wondering if
>>>> it possible
>>>> for query planner to push down the order by clause at the datanode and
>>>> then perform
>>>>  sort-merge aggregate at the coordinator. Surely, that would be a
>>>> better query plan.
>>>>
>>>>  We have tried enable_sort=off etc. but that doesn't work.
>>>>
>>>>  Thanks.
>>>>  Sandeep
>>>>
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> October Webinars: Code for Performance
>>>> Free Intel webinars can help you accelerate application performance.
>>>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the
>>>> most from
>>>> the latest Intel processors and coprocessors. See abstracts and
>>>> register >
>>>>
>>>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>>> _______________________________________________
>>>> Postgres-xc-general mailing list
>>>> Pos...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>>>
>>>>
>>>
>>>
>>> --
>>> Best Wishes,
>>> Ashutosh Bapat
>>> EnterpriseDB Corporation
>>> The Postgres Database Company
>>>
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

How about a plan where the datanodes perform the sort as well and the
coordinator performs a sorted merge?

Are such plans not part of the query planner?

-Sandeep

On 10/17/2013 07:04 AM, Ashutosh Bapat wrote:
> There is GROUP BY clause that needs to be evaluated before the result 
> can be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.
>
>
> On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta 
> <gup...@gm... <mailto:gup...@gm...>> wrote:
>
>     Hi Ashutosh,
>
>
>       Attached below is the query and the corresponding query plan. I
>     am using version 1.1.
>     Thanks for taking a look at this.
>
>     -Sandeep
>
>     SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>
>
>
>
>                                         QUERY PLAN (Coordinator)
>
>       Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>         Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>         Sort Key: a.exposed_time
>         ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>               Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>               ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>                     Output: a.exposed_time, (count(a.pid))
>                     Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>                     Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
>     (9 rows)
>
>                                           QUERY PLAN (Datanode)
>
>       GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>         Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>         ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>               Output: a.exposed_time, a.pid
>               ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>                     Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>                     Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>               ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>                     Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>                     Index Cond: (d.pid = a.pid)
>                     Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
>     (11 rows)
>
>
>
>
>     On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat
>     <ash...@en...
>     <mailto:ash...@en...>> wrote:
>
>         Sandeep,
>         It would be nice if you mention the version of XC in your
>         mail. Sort push down is available from 1.1 onwards. If you do
>         not see sort getting pushed down in 1.1, please report
>         detailed definitions of the tables, query and the EXPLAIN output.
>
>
>         On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta
>         <gup...@gm... <mailto:gup...@gm...>> wrote:
>
>             Hi,
>
>              In an another query that requires the result to be
>             aggregated and ordered by a field (lets say timeo)
>             the query planner currently pulls the results and then
>             performs a sort with hash aggregate.
>
>             The table at the datanodes are clustered by timeo. I was
>             wondering if it possible
>             for query planner to push down the order by clause at the
>             datanode and then perform
>             sort-merge aggregate at the coordinator. Surely, that
>             would be a better query plan.
>
>             We have tried enable_sort=off etc. but that doesn't work.
>
>             Thanks.
>             Sandeep
>
>
>
>             ------------------------------------------------------------------------------
>             October Webinars: Code for Performance
>             Free Intel webinars can help you accelerate application
>             performance.
>             Explore tips for MPI, OpenMP, advanced profiling, and
>             more. Get the most from
>             the latest Intel processors and coprocessors. See
>             abstracts and register >
>             http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>             _______________________________________________
>             Postgres-xc-general mailing list
>             Pos...@li...
>             <mailto:Pos...@li...>
>             https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>
>
>
>
>         -- 
>         Best Wishes,
>         Ashutosh Bapat
>         EnterpriseDB Corporation
>         The Postgres Database Company
>
>
>
>
>
> -- 
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company

There is GROUP BY clause that needs to be evaluated before the result can
be ordered. Thus GROUP BY is sent to the datanode but not ORDER BY.


On Thu, Oct 17, 2013 at 4:31 PM, Sandeep Gupta <gup...@gm...>wrote:

> Hi Ashutosh,
>
>
>   Attached below is the query and the corresponding query plan. I am using
> version 1.1.
> Thanks for taking a look at this.
>
> -Sandeep
>
> SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;
>
>
>                                    QUERY PLAN (Coordinator)
>
>  Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
>    Output: a.exposed_time, (round(((count((count(a.pid))) / 10))::double precision))
>    Sort Key: a.exposed_time
>    ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
>          Output: a.exposed_time, round(((count((count(a.pid))) / 10))::double precision)
>          ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"  (cost=0.00..0.00 rows=1000 width=8)
>                Output: a.exposed_time, (count(a.pid))
>                Node/s: datanode1, datanode10, datanode11, datanode12, datanode13, datanode14, datanode15, datanode16, datanode2, datanode3, datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
>                Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 = r.a_2) GROUP BY 1
> (9 rows)
>
>                                      QUERY PLAN (Datanode)
>
>  GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
>    Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
>    ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
>          Output: a.exposed_time, a.pid
>          ->  Index Scan using et_ses on public.ses_vt_20130805_xc a  (cost=0.00..7283.10 rows=129583 width=8)
>                Output: a.pid, a.rep, a.exposed_time, a.infectious_time, a.recovered_time
>                Index Cond: ((a.exposed_time >= 4667) AND (a.exposed_time <= 5031))
>          ->  Index Scan using pid_demo on public.vt_demography_info_xc d  (cost=0.00..0.30 rows=1 width=4)
>                Output: d.pid, d.hid, d.age, d.gender, d.zipode, d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
>                Index Cond: (d.pid = a.pid)
>                Filter: ((d.age >= 5) AND (d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender = 1))
> (11 rows)
>
>
>
>
> On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
> ash...@en...> wrote:
>
>> Sandeep,
>> It would be nice if you mention the version of XC in your mail. Sort push
>> down is available from 1.1 onwards. If you do not see sort getting pushed
>> down in 1.1, please report detailed definitions of the tables, query and
>> the EXPLAIN output.
>>
>>
>> On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <gup...@gm...>wrote:
>>
>>> Hi,
>>>
>>>  In an another query that requires the result to be aggregated and
>>> ordered by a field (lets say timeo)
>>> the query planner currently pulls  the results and then performs a sort
>>> with hash aggregate.
>>>
>>> The table at the datanodes are clustered by timeo. I was wondering if it
>>> possible
>>> for query planner to push down the order by clause at the datanode and
>>> then perform
>>> sort-merge aggregate at the coordinator. Surely, that would be a better
>>> query plan.
>>>
>>> We have tried enable_sort=off etc. but that doesn't work.
>>>
>>> Thanks.
>>> Sandeep
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> October Webinars: Code for Performance
>>> Free Intel webinars can help you accelerate application performance.
>>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
>>> from
>>> the latest Intel processors and coprocessors. See abstracts and register
>>> >
>>>
>>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>>> _______________________________________________
>>> Postgres-xc-general mailing list
>>> Pos...@li...
>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>>
>>>
>>
>>
>> --
>> Best Wishes,
>> Ashutosh Bapat
>> EnterpriseDB Corporation
>> The Postgres Database Company
>>
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

Hi Ashutosh,


  Attached below is the query and the corresponding query plan. I am using
version 1.1.
Thanks for taking a look at this.

-Sandeep

SELECT exposed_time effectedDate, ROUND(COUNT(a.pid)/10) COUNT FROM
public.vt_demography_info_xc d, public.ses_vt_20130805_xc a WHERE
d.pid=a.pid AND d.countyid='50015' AND d.age BETWEEN 5 AND 18 AND
d.gender=1 GROUP BY exposed_time ORDER BY exposed_time;


                                   QUERY PLAN (Coordinator)

 Sort  (cost=10000000005.03..10000000005.03 rows=1 width=8)
   Output: a.exposed_time, (round(((count((count(a.pid))) /
10))::double precision))
   Sort Key: a.exposed_time
   ->  HashAggregate  (cost=5.00..5.02 rows=1 width=8)
         Output: a.exposed_time, round(((count((count(a.pid))) /
10))::double precision)
         ->  Data Node Scan on "__REMOTE_GROUP_QUERY__"
(cost=0.00..0.00 rows=1000 width=8)
               Output: a.exposed_time, (count(a.pid))
               Node/s: datanode1, datanode10, datanode11, datanode12,
datanode13, datanode14, datanode15, datanode16, datanode2, datanode3,
datanode4, datanode5, datanode6, datanode7, datanode8, datanode9
               Remote query: SELECT r.a_1, count(r.a_2) FROM ((SELECT
d.pid FROM ONLY public.vt_demography_info_xc d WHERE ((d.age >= 5) AND
(d.age <= 18) AND ((d.countyid)::text = '50015'::text) AND (d.gender =
1))) l(a_1) JOIN (SELECT a.exposed_time, a.pid FROM ONLY
public.ses_vt_20130805_xc a WHERE ((a.exposed_time >= 4667) AND
(a.exposed_time <= 5031))) r(a_1, a_2) ON (true)) WHERE (l.a_1 =
r.a_2) GROUP BY 1
(9 rows)

                                     QUERY PLAN (Datanode)

 GroupAggregate  (cost=0.00..47862.29 rows=225 width=8)
   Output: a.exposed_time, round(((count(a.pid) / 10))::double precision)
   ->  Nested Loop  (cost=0.00..47856.05 rows=460 width=8)
         Output: a.exposed_time, a.pid
         ->  Index Scan using et_ses on public.ses_vt_20130805_xc a
(cost=0.00..7283.10 rows=129583 width=8)
               Output: a.pid, a.rep, a.exposed_time,
a.infectious_time, a.recovered_time
               Index Cond: ((a.exposed_time >= 4667) AND
(a.exposed_time <= 5031))
         ->  Index Scan using pid_demo on public.vt_demography_info_xc
d  (cost=0.00..0.30 rows=1 width=4)
               Output: d.pid, d.hid, d.age, d.gender, d.zipode,
d.blockgroupid, d.longitude, d.lattitude, d.county, d.countyid
               Index Cond: (d.pid = a.pid)
               Filter: ((d.age >= 5) AND (d.age <= 18) AND
((d.countyid)::text = '50015'::text) AND (d.gender = 1))
(11 rows)




On Thu, Oct 17, 2013 at 12:12 AM, Ashutosh Bapat <
ash...@en...> wrote:

> Sandeep,
> It would be nice if you mention the version of XC in your mail. Sort push
> down is available from 1.1 onwards. If you do not see sort getting pushed
> down in 1.1, please report detailed definitions of the tables, query and
> the EXPLAIN output.
>
>
> On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <gup...@gm...>wrote:
>
>> Hi,
>>
>>  In an another query that requires the result to be aggregated and
>> ordered by a field (lets say timeo)
>> the query planner currently pulls  the results and then performs a sort
>> with hash aggregate.
>>
>> The table at the datanodes are clustered by timeo. I was wondering if it
>> possible
>> for query planner to push down the order by clause at the datanode and
>> then perform
>> sort-merge aggregate at the coordinator. Surely, that would be a better
>> query plan.
>>
>> We have tried enable_sort=off etc. but that doesn't work.
>>
>> Thanks.
>> Sandeep
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> October Webinars: Code for Performance
>> Free Intel webinars can help you accelerate application performance.
>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
>> from
>> the latest Intel processors and coprocessors. See abstracts and register >
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Postgres-xc-general mailing list
>> Pos...@li...
>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>

Yes, foreign data is not a part of XC cluster.   Foreign data's transaction management is separate from the cluster and we cannot enforce data integrity.   Even though we support FDW, it is just foreign data, not a part of XC cluster.   To make PostgreSQL data as a part of XC cluster, PG needs to accept GXID and snapshot from XC, as well as sequence, if it is shared with other tables in XC.

It does not sound simple because in this case, PostgreSQL database is not autonomous and cannot operate by its own.

Regards;
---
Koichi Suzuki

On 2013/10/17, at 14:52, Amit Khandekar <ami...@en...<mailto:ami...@en...>>
 wrote:

I think Aris's expectation is that by using FDW support we can create a cluster of heterogeneous nodes. This is not going to happen just by supporting  foreign data wrappers and foreign tables.

Note that allowing a foreign table/server to be created in Postgres-XC means a foreign table will be created on a machine which is completely outside the Postgres-XC cluster. Making that machine a part of the cluster is a different thing, and that does not require FDWs.


On 17 October 2013 10:48, 鈴木 幸市 <koichi@intellilink..co.jp<mailto:ko...@in...>> wrote:
So far, CREATE FOREIGN DATA WRAPPER, CREATE SERVER and CREATE USER MAPPING are blocked.   As Michael suggests, yes, it would be nice to connect to foreign data through coordinators or even datanodes.

It's welcome if more people are involved in the test, not just development.   Contribution of the code is more than welcome.

Unfortunately, nobody dis these work mainly due to the resource.   it will be wonderful if anybody can join and contribute the code.   There are not reason that XC doesn't have to support FDW.

Best;
---
Koichi Suzuki

On 2013/10/17, at 13:11, Michael Paquier <mic...@gm...<mailto:mic...@gm...>> wrote:

> On Thu, Oct 17, 2013 at 12:43 PM, Aris Setyawan <ari...@gm...<mailto:ari...@gm...>> wrote:
>> Hi,
>>
>> Can XC be used with [write-able] foreign-data wrapper? What I mean
>> here are push down optimization and data distribution.
> XC does not support itself fdw, but I don't see why there would be
> problems to have a Postgres server with a postgres_fdw connect to
> Coordinators for read/write operations or even Datanodes for read-only
> operations as the communication interface is the same as vanilla. Take
> care to use at least XC 1.1~ for the latter though.
>
>> I imagine, If yes, we can have a cluster of not just postgresql node.
>> But we can have oracle or mysql or redis or unlimited cluster.
> Yep. Supporting FDW in XC would be fun as well. Patches welcome.
> --
> Michael
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...<mailto:Pos...@li...>
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>


------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
_______________________________________________
Postgres-xc-general mailing list
Pos...@li...<mailto:Pos...@li...>
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general

I think Aris's expectation is that by using FDW support we can create a
cluster of heterogeneous nodes. This is not going to happen just by
supporting  foreign data wrappers and foreign tables.

Note that allowing a foreign table/server to be created in Postgres-XC
means a foreign table will be created on a machine which is completely
outside the Postgres-XC cluster. Making that machine a part of the cluster
is a different thing, and that does not require FDWs.


On 17 October 2013 10:48, 鈴木 幸市 <ko...@in...> wrote:

> So far, CREATE FOREIGN DATA WRAPPER, CREATE SERVER and CREATE USER MAPPING
> are blocked.   As Michael suggests, yes, it would be nice to connect to
> foreign data through coordinators or even datanodes.
>
> It's welcome if more people are involved in the test, not just
> development.   Contribution of the code is more than welcome.
>
> Unfortunately, nobody dis these work mainly due to the resource.   it will
> be wonderful if anybody can join and contribute the code.   There are not
> reason that XC doesn't have to support FDW.
>
> Best;
> ---
> Koichi Suzuki
>
> On 2013/10/17, at 13:11, Michael Paquier <mic...@gm...>
> wrote:
>
> > On Thu, Oct 17, 2013 at 12:43 PM, Aris Setyawan <ari...@gm...>
> wrote:
> >> Hi,
> >>
> >> Can XC be used with [write-able] foreign-data wrapper? What I mean
> >> here are push down optimization and data distribution.
> > XC does not support itself fdw, but I don't see why there would be
> > problems to have a Postgres server with a postgres_fdw connect to
> > Coordinators for read/write operations or even Datanodes for read-only
> > operations as the communication interface is the same as vanilla. Take
> > care to use at least XC 1.1~ for the latter though.
> >
> >> I imagine, If yes, we can have a cluster of not just postgresql node.
> >> But we can have oracle or mysql or redis or unlimited cluster.
> > Yep. Supporting FDW in XC would be fun as well. Patches welcome.
> > --
> > Michael
> >
> >
> ------------------------------------------------------------------------------
> > October Webinars: Code for Performance
> > Free Intel webinars can help you accelerate application performance.
> > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> > the latest Intel processors and coprocessors. See abstracts and register
> >
> >
> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
> > _______________________________________________
> > Postgres-xc-general mailing list
> > Pos...@li...
> > https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
> >
>
>
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>

So far, CREATE FOREIGN DATA WRAPPER, CREATE SERVER and CREATE USER MAPPING are blocked.   As Michael suggests, yes, it would be nice to connect to foreign data through coordinators or even datanodes. 

It's welcome if more people are involved in the test, not just development.   Contribution of the code is more than welcome.

Unfortunately, nobody dis these work mainly due to the resource.   it will be wonderful if anybody can join and contribute the code.   There are not reason that XC doesn't have to support FDW.

Best;
---
Koichi Suzuki

On 2013/10/17, at 13:11, Michael Paquier <mic...@gm...> wrote:

> On Thu, Oct 17, 2013 at 12:43 PM, Aris Setyawan <ari...@gm...> wrote:
>> Hi,
>> 
>> Can XC be used with [write-able] foreign-data wrapper? What I mean
>> here are push down optimization and data distribution.
> XC does not support itself fdw, but I don't see why there would be
> problems to have a Postgres server with a postgres_fdw connect to
> Coordinators for read/write operations or even Datanodes for read-only
> operations as the communication interface is the same as vanilla. Take
> care to use at least XC 1.1~ for the latter though.
> 
>> I imagine, If yes, we can have a cluster of not just postgresql node.
>> But we can have oracle or mysql or redis or unlimited cluster.
> Yep. Supporting FDW in XC would be fun as well. Patches welcome.
> -- 
> Michael
> 
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
> 

Sandeep,
It would be nice if you mention the version of XC in your mail. Sort push
down is available from 1.1 onwards. If you do not see sort getting pushed
down in 1.1, please report detailed definitions of the tables, query and
the EXPLAIN output.


On Thu, Oct 17, 2013 at 1:09 AM, Sandeep Gupta <gup...@gm...>wrote:

> Hi,
>
>  In an another query that requires the result to be aggregated and ordered
> by a field (lets say timeo)
> the query planner currently pulls  the results and then performs a sort
> with hash aggregate.
>
> The table at the datanodes are clustered by timeo. I was wondering if it
> possible
> for query planner to push down the order by clause at the datanode and
> then perform
> sort-merge aggregate at the coordinator. Surely, that would be a better
> query plan.
>
> We have tried enable_sort=off etc. but that doesn't work.
>
> Thanks.
> Sandeep
>
>
>
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

On Thu, Oct 17, 2013 at 12:43 PM, Aris Setyawan <ari...@gm...> wrote:
> Hi,
>
> Can XC be used with [write-able] foreign-data wrapper? What I mean
> here are push down optimization and data distribution.
XC does not support itself fdw, but I don't see why there would be
problems to have a Postgres server with a postgres_fdw connect to
Coordinators for read/write operations or even Datanodes for read-only
operations as the communication interface is the same as vanilla. Take
care to use at least XC 1.1~ for the latter though.

> I imagine, If yes, we can have a cluster of not just postgresql node.
> But we can have oracle or mysql or redis or unlimited cluster.
Yep. Supporting FDW in XC would be fun as well. Patches welcome.
-- 
Michael

Hi,

Can XC be used with [write-able] foreign-data wrapper? What I mean
here are push down optimization and data distribution.

I imagine, If yes, we can have a cluster of not just postgresql node.
But we can have oracle or mysql or redis or unlimited cluster.

-Aris

Hi,

 In an another query that requires the result to be aggregated and ordered
by a field (lets say timeo)
the query planner currently pulls  the results and then performs a sort
with hash aggregate.

The table at the datanodes are clustered by timeo. I was wondering if it
possible
for query planner to push down the order by clause at the datanode and then
perform
sort-merge aggregate at the coordinator. Surely, that would be a better
query plan.

We have tried enable_sort=off etc. but that doesn't work.

Thanks.
Sandeep

Sorry I did not respond for a while.   Please take a look at my comment
inline.

Regards;
---
Koichi Suzuki


2013/10/8 Yehezkel Horowitz <hor...@ch...>

> >> My goal - I have an application that needs SQL DB and must always be
> >> up (I have a backup machine for this purpose).
> >Have you thought about PostgreSQL itself for your solution. Is there any
> reason you'd need XC? Do you have an amount of data that >forces you to use
> multi-master architecture or perhaps PG itself could handle it?
>
> I need multi-master capability, as clients might connect to both machines
> at the same time; Yes - my tables will be replicated.
>
> >Yep, this is doable. If all your data is replicated you would be able to
> do that. However you need to keep in mind that you will not be able to
> write new data to node B if node A is not accessible. If you data is
> replicated and you need to update a table, both nodes need to work.
>
> This is a surprise for me, this wasn't clear in the documentation I read
> nor at some PG-XC presentations I looked at in the internet.
> Isn't this point one of the conditions for High-Availability of DB -
> allowing work to continue even if one of the machines failed?
>

Postgres-XC assumes any table may be replicated or distributed so XC does
not have an operation interface assuming all the tables are replicated.
It always assumes there could be some table distributed, some replicated.

On the other hand, Postgres-XC's most important feature is to maintain
cluster-wide data integrity.   XC's replication is for HA, but to provide
scalability by proxying as many statement to local datanode and increase
parallelism.

So, when you issue a DML against a replicated table, Postgres-XC tries to
propagate it to all the nodes where it is defined over.    If any node is
not available, Postgres-XC determines it cannot maintain cluster-wide data
integrity.

We provide a couple of means to deal with this.

1. ALTER TABLE to change table's replication.   You can delete any node.
Because this change should go to any other nodes for cluster-wide data
integrity, you should have all the datanodes working.

2. Configure slaves for each master.   When one of them fails, it can be
failed over by its slave.   Typically, you can configure slaves at other
datanode's server each other.  After failover occurs (you may want to
integrate with automatic failover system such as Pacemaker and
Corosync/Heartbeat) and you feel it's not needed any longer, you can issue
ALTER TEABLE to delete failed node your cluster, issue DROP NODE as well,
and then stop the slave and release its resource.


> >Or if you want B to be still writable, you could update the node
> information inside it, make it workable alone, and when server A is up
> again recreate a new XC node from scratch and add it again to the cluster.
>
> What is the correct procedure for doing that? Is there a pgxc_ctl commands
> for doing that?
>

Hope the above helps.


>
> >> My questions:
> >>
> >> 1.       In your docs, you always put the GTM in dedicated machine.
> >> a.       Is this a requirement, just an easy to understand topology or
> best
> >> practice?
> >GTM consumes a certain amount of CPU and does not need much RAM, while
> for your nodes you might prioritize the opposite.
> >> b.      In case of best practice, what is the expected penalty in case
> the
> >> GTM is deployed on the same machine with coordinator and datanode?
> >CPU resource consumption and reduction of performance if your queries
> need some CPU with for example internal sort operations among other things.
> O.K  got it; For now I'm trying to make it work, afterwards I'll take care
> for make it work faster.
>
> >> 2.       What should I do after Machine A is back to life if I want:
> >> a.       Make it act as a new slave?
> >> b.      Make it become the master again?
> >There is no principle of master/slave in XC like in Postgres (well you
> could create a slave node for an individual Coordinator/Datanode). >But
> basically in your configuration machine A and B have the same state.
> >Only GTM is a slave.
>
> Sorry, I meant in the context of GTM - how should I make MachineA a new
> GTM-slave or make it a GTM-master again?
>

You need to configure gtm_proxy for this purpose.   Gtm_ctl provides
failover option for gtm slave to be the new gtm master.   It also provides
reconnect option for gtm_proxy to connect to the new gtm master.
 Pgxc_ctl provides this as corresponding commands.   Please take a look at
http://postgres-xc.sourceforge.net/docs/1_1/pgxc-ctl.html


>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>

On Mon, Oct 14, 2013 at 4:05 PM, admin <ad...@75...> wrote:
> Hello, I'm a newbie and tring to evaluate pgxc for a project.
> I need to say pgxc is an interesting solution of cluster.
> In the evaluation I found some questions, if somebody can answer it I
> will be very glad.
>
> Environment:
> pgxc 1.1, centos 6 32bit in virtualbox
> ServerA:
> gtm
> coord_01
> datanode_01
> ServerB:
> datanode_02
>
> Question:
> a. Is connection secure between coordinator with datanode, or
> coordinator with gtm?
> I tried if datanode only allow password authentication it will not
> accessable from coordinator,
> and gtm have no such config like pg_hba.conf
No there is no option to use SSL between those connections due to the
internal connection pooling implementation. This is a TODO item.

> b. What is the warning "Do not have a GTM snapshot available" mean?
This warning means that you got a problem when trying to request a
global snapshot when issuing a query on a node. This was for example
reproducible easily by running a query directly on a Datanode without
going though a Coordinator for example.

> c. If gtm_proxy go after gtm, it will direct exit, is it a problem?
> Example if I'm going to write a loader, step cannot be
> gtm_ctl -Z gtm -D gtm start
> gtm_ctl -Z gtm_proxy -D gtm_proxy start
> but must be
> gtm_ctl -Z gtm -D gtm start
> [wait the port use by gtm open]
> gtm_ctl -Z gtm_proxy -D gtm_proxy start
Do you mean if gtm_proxy is started before gtm? Perhaps others
(Pavan?) will correct me, but isn't the gtm_proxy to not exit directly
and wait for a port open?

> d. Check only one node is primary seems have a bug, example:
> select * from pgxc_node;
> "coord_01";"C";5432;"localhost";f;f;1975432854
> "datanode_02";"D";15432;"192.168.8.184";f;f;-1414354208
> "datanode_01";"D";15432;"192.168.8.183";t;t;-1746477557
>
> select pgxc_pool_reload();
> (success)
> alter node datanode_01 with (port=15433);
> ERROR: PGXC node datanode_01: two nodes cannot be primary
> alter node datanode_01 with (primary=true);
> ERROR: PGXC node datanode_01: two nodes cannot be primary
> alter node datanode_01 with (primary=false);
> (success)
Definitely looks like a bug as you are describing it.

> These command all done in the same session open from pgadmin.
> e. I cant set primary key to a table if remote node added, is it not
> supported yet?
> alter table test add primary key (id);
> ERROR: Cannot create index whose evaluation cannot be enforced to remote
> nodes
You need either to make the table test replicated, or hashed using id
as a key in this case. AFAIK pgadmin does not provide support for the
XC-specific query extensions, but you can always run raw SQLs.

> f. I cant modify value of sequence, is it not supported yet?
> select setval('public.test_01_id_seq', 123, true);
> ERROR: GTM error, could not obtain sequence value
This should be supported.

> g. I found some time field name and its type will exchange,
> like "name with type text" some time will trun to "text with type name",
> and type "name" is not valid.
> I have no reproduction way yet but is it a known bug?
Can you provide a test case here?

> h. I found some time sequence from serial will broken then i can't
> insert any data,
> select nextval('public.test_01_id_seq');
> ERROR:
> Status: XX00
> I can upload database files if you need.
Not sure I am following you here.

> i. Is there some way to get what nodes are using with the table?
> These command can manage nodes with table
> alter table tablename add node ...
> alter table tablename delete node ...
> alter table tablename to node ...
> but I dont known what command can list what nodes are using with the table.
pgxc_class contains a list of node OIDS to know where the table data
is located referring to the nodes in pgxc_node.

> j. Is this project ready for production use?
Some do.
-- 
Michael

Hello, I'm a newbie and tring to evaluate pgxc for a project.
I need to say pgxc is an interesting solution of cluster.
In the evaluation I found some questions, if somebody can answer it I
will be very glad.

Environment:
pgxc 1.1, centos 6 32bit in virtualbox
ServerA:
gtm
coord_01
datanode_01
ServerB:
datanode_02

Question:
a. Is connection secure between coordinator with datanode, or
coordinator with gtm?
I tried if datanode only allow password authentication it will not
accessable from coordinator,
and gtm have no such config like pg_hda.conf
b. What is the warning "Do not have a GTM snapshot available" mean?
c. If gtm_proxy go after gtm, it will direct exit, is it a problem?
Example if I'm going to write a loader, step cannot be
gtm_ctl -Z gtm -D gtm start
gtm_ctl -Z gtm_proxy -D gtm_proxy start
but must be
gtm_ctl -Z gtm -D gtm start
[wait the port use by gtm open]
gtm_ctl -Z gtm_proxy -D gtm_proxy start
d. Check only one node is primary seems have a bug, example:
select * from pgxc_node;
"coord_01";"C";5432;"localhost";f;f;1975432854
"datanode_02";"D";15432;"192.168.8.184";f;f;-1414354208
"datanode_01";"D";15432;"192.168.8.183";t;t;-1746477557

select pgxc_pool_reload();
(success)

alter node datanode_01 with (port=15433);
ERROR: PGXC node datanode_01: two nodes cannot be primary

alter node datanode_01 with (primary=true);
ERROR: PGXC node datanode_01: two nodes cannot be primary

alter node datanode_01 with (primary=false);
(success)

select pgxc_pool_reload();
(success)

select * from pgxc_node;
"coord_01";"C";5432;"localhost";f;f;1975432854
"datanode_02";"D";15432;"192.168.8.184";f;f;-1414354208
"datanode_01";"D";15432;"192.168.8.183";f;t;-1746477557

alter node datanode_01 with (primary=true);
ERROR: PGXC node datanode_01: two nodes cannot be primary

These command all done in the same session open from pgadmin.
e. I cant set primary key to a table if remote node added, is it not
supported yet?
alter table test add primary key (id);
ERROR: Cannot create index whose evaluation cannot be enforced to remote
nodes
f. I cant modify value of sequence, is it not supported yet?
select setval('public.test_01_id_seq', 123, true);
ERROR: GTM error, could not obtain sequence value
g. I found some time field name and its type will exchange,
like "name with type text" some time will trun to "text with type name",
and type "name" is not valid.
I have no reproduction way yet but is it a known bug?
h. I found some time sequence from serial will broken then i can't
insert any data,
select nextval('public.test_01_id_seq');
ERROR:
Status: XX00
I can upload database files if you need.
i. Is there some way to get what nodes are using with the table?
These command can manage nodes with table
alter table tablename add node ...
alter table tablename delete node ...
alter table tablename to node ...
but I dont known what command can list what nodes are using with the table.
j. Is this project ready for production use?

2nd try.
Can you please answer my questions below?

TIA

Yehezkel Horowitz

-----Original Message-----
From: Yehezkel Horowitz 
Sent: Tuesday, October 08, 2013 2:30 PM
To: 'Michael Paquier'; <pos...@li...>
Subject: RE: [Postgres-xc-general] Some questions about postgres-XC

>> My goal - I have an application that needs SQL DB and must always be 
>> up (I have a backup machine for this purpose).
>Have you thought about PostgreSQL itself for your solution. Is there any reason you'd need XC? Do you have an amount of data that >forces you to use multi-master architecture or perhaps PG itself could handle it?

I need multi-master capability, as clients might connect to both machines at the same time; Yes - my tables will be replicated.

>Yep, this is doable. If all your data is replicated you would be able to do that. However you need to keep in mind that you will not be able to write new data to node B if node A is not accessible. If you data is replicated and you need to update a table, both nodes need to work.

This is a surprise for me, this wasn't clear in the documentation I read nor at some PG-XC presentations I looked at in the internet.
Isn't this point one of the conditions for High-Availability of DB - allowing work to continue even if one of the machines failed?

>Or if you want B to be still writable, you could update the node information inside it, make it workable alone, and when server A is up again recreate a new XC node from scratch and add it again to the cluster.

What is the correct procedure for doing that? Is there a pgxc_ctl commands for doing that?

>> My questions:
>>
>> 1.       In your docs, you always put the GTM in dedicated machine.
>> a.       Is this a requirement, just an easy to understand topology or best
>> practice?
>GTM consumes a certain amount of CPU and does not need much RAM, while for your nodes you might prioritize the opposite.
>> b.      In case of best practice, what is the expected penalty in case the
>> GTM is deployed on the same machine with coordinator and datanode?
>CPU resource consumption and reduction of performance if your queries need some CPU with for example internal sort operations among other things.
O.K  got it; For now I'm trying to make it work, afterwards I'll take care for make it work faster.

>> 2.       What should I do after Machine A is back to life if I want:
>> a.       Make it act as a new slave?
>> b.      Make it become the master again?
>There is no principle of master/slave in XC like in Postgres (well you could create a slave node for an individual Coordinator/Datanode). >But basically in your configuration machine A and B have the same state.
>Only GTM is a slave.

Sorry, I meant in the context of GTM - how should I make MachineA a new GTM-slave or make it a GTM-master again?

On 10/5/2013 5:19 PM, Michael Paquier wrote:
> On Sat, Oct 5, 2013 at 9:00 PM, Stefan Lekov <ar...@er...> wrote:
>> Hello, I'm new to the Postgres-XC project. In fact I am still
>> considering if I should install it in order to try it as a
>> replacement of my current database clusters (those are based around
>> MySQL and its binary_log based replication).
> Have you considered PostgreSQL as a potential solution before
> Postgres-XC. Why do you especially need XC?
I have used PosgreSQL in the past, I am using it at the moment (for 
other projects) and I'd like to continue using in the future. My current 
requirements are including having a multi-master replicated database 
cluster. These requirements are related to redundancy and possible 
scalability. While one PostgreSQL server will cope with any load that I 
can throw at it for the near future, that might not be the case in about 
an year or two. As for the redundancy part - I am familiar with 
PostgreSQL capabilities of a warm standby server however I am looking 
for something more robust.

Because of the requirement of "multi-master", I am investigating 
Postgres-XC and pgpool2 capabilities to deliver such system. I can 
migrate to a single PostgreSQL server, however I am not really keen on 
solving the replication dilemma on-the-fly when the system is already 
running with Postgres - I prefer having something that is already 
working as expected right from the start.

>> Before actually starting the installation of postgres-xc I would like
>> to know what is the procedure for restarting nodes. I have already
>> read a few documents/mails regarding restoring or resyncing a failed
>> datanode, however these documents does not answer my simple question:
>> What should be the procedure for rebooting servers? For example I
>> have a kernel updated pending (due to security reasons) - I'm
>> installing the new kernel, but I have to reboot all machine.
>> Theoretically all nodes (both coordinators and datanodes) are working
>> on different physical servers or VMes. In a perfect scenario I would
>> like to keep the system in production while I am restarting the
>> servers one by one. However I am not sure what would be the effect of
>> rebooting servers one by one.
> If a node is restarted or facing an outage, all the transactions it
> needs to be involved in will simply fail. In the case of Coordinator,
> this has effect only for DDL. For Datanodes, this has effect as well
> for DDL, but also for DML and SELECT of the node is needed for the
> transaction.
There would be no DDL during these operations. I can limit the queries 
to DML only.
>> For purpose of example let me have four datanodes: A,B,C,D All
>> servers are synced and are operating as expected. 1) Upgrade A,
>> reboot A 2) INSERT/UPDATE/DELETE queries 3) A boots up and is
>> successfully started 4) INSERT/UPDATE/DELETE queries 5) Upgrade B,
>> reboot B ... ... As for the "Coordinators" nodes. How are those
>> affected by temporary stopping and restarting the postgres-xc related
>> services. What should be the load balancer in front of these servers
>> in order to be able to both load-balance and fail-over if one of the
>> Coordinators is offline either due to failed server or due to
>> rebooting servers.
> DDLs won't work. Applications will use one access point. In this case
> no problems for your application, connect to the other Coordinators to
> execute queries as long as they are not DDLs.
What system, application or method would you recommend for performing 
the load-balance/fail-over of connections to the Coordinators.
>> I have no problem with relatively heavy operation of full restore of
>> a datanode in event of failed server. Such restoration operation can
>> be properly scheduled and executed, however I am interested how would
>> postgres-xc react to simple scenarioa simple operation of restarting
>> a server due to whatever reasons should
> As mentioned above, transactions that will need it will simply fail.
> You could always failover a slave for the outage period if necessary.
Correct me if I'm wrong: All data (read databases, schema, tables, etc) 
would be replicated to all datanodes. So before host A goes down all 
servers would have the same dataset. This way no transaction should fail 
due to the missing datanode A. While A has been booting up several 
transactions have passed (since such restart is an operation I can 
schedule, I'm doing that during time when we have low to no load on our 
systems, thus the transaction count is relatively low). My question is 
how to bring A back to having "the same dataset" as the rest of the 
datanodes before I can continue with the next host/datanode?

Regards,
Stefan Lekov

------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Postgres-xc-general mailing list
Pos...@li...
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general

On Wed, Oct 9, 2013 at 9:54 PM, Hector M. Jacas <hec...@et...>wrote:

>  Hi all,
>
> First I must apologize because obviously I was reading the document
> oriented to bash version of pgxc_ctl.
>
> In the documentation of the binary version there is no reference to these
> facilities (dropdb/dropuser). This is the one I should have read .
>
> My mistake and my apologies .
>
> I beg your patience and condescension when from the user role of this
> great project I take the liberty to comment the answer to my post .
>
> I do not share the view of Mr. Ashutosh Bapat when he says " is not an
> interface pgxc_ctl for dropping database or user. " reason: "It's just a
> cluster management utility "
>
> I think there is an inconsistency in that statement because the same
> reason for not including dropdb and dropuser commands are perfectly valid
> createdb and createuser .
>

There is difference between what is supported as a requirement and what is
supported because it fits well in the utility. You may compare pgxc_ctl
with pg_ctl, which basically allow controlling the life of server. pgxc_ctl
being made for XC, has to support life of a cluster and allows controlling
individual server. On top, it allows creating a cluster, (which is not
required in pg_ctl, initdb does it). This particular functionality needs
Createdb and Createuser, so does it support those. But a user should not
look at pgxc_ctl for managing individual databases. The server is more than
capable of doing it and that functionality can be accessed through
connectors or utilities like create* or drop*.


>
> You as project developers ( or contribution ) decide the philosophy with
> which your product works and I in my role as user I should be able to
> reconcile my working methods with the philosophy of the tools I have
> selected.
>
> POSTGRESXC is a great project because it solves big problems.
>
> PGXC_CTL is another great project because it simplifies the deployment and
> management of postgresxc and if you add shortcuts to frequently used
> commands (and perhaps, some security features) this project could become a
> kind of Central Command for POSTGRESXC .
>
> Thank you very much for your answers ,
>
> Hector M. Jacas
>
>
>
>
> On 10/09/2013 12:12 AM, Ashutosh Bapat wrote:
>
>  Hector,
>  AFAIK, pgxc_ctl is not an interface for dropping database or user. It's
> just a cluster management utility. You should use corresponding binaries or
> SQL commands for that purpose.
>
>
> On Tue, Oct 8, 2013 at 9:32 PM, Hector M. Jacas <hec...@et...>wrote:
>
>>
>> Hi all,
>>
>> Among the features described in:
>> https://github.com/koichi-szk/PGXC-Tools/blob/master/pgxc_ctl/manual.txtis deleting the databases (Dropdb) and users (Dropuser) and when I try make
>> use of these commands pgxc_ctl answers: command not found
>>
>> PGXC Createdb testdb
>> Selected coord2.
>> PGXC Dropdb testdb
>> sh: Dropdb: command not found
>> PGXC Createuser usertest1
>> Selected coord1.
>> PGXC Dropuser usertest1
>> sh: Dropuser: command not found
>> PGXC
>>
>> Carefully review the source code and found that in the folder:
>> postgres-xc/contrib/pgxc_ctl , there is a file (do_command.c) in which
>> reference is made and performed the execution of Createdb (line 2339) and
>> Createuser (line 2369).
>>
>> In this file there is no reference whatsoever to Dropdb or Dropuser .
>>
>> There is another file (in the same directory) called: pgxc_ctl.bash, in
>> which reference is made and run the corresponding command to Createdb,
>> Dropdb, Createuser and Dropuser.
>>
>> Do not remember reading during pgxc compliacion and deployment (or
>> pgxc_ctl in the area of contributions ) anything regarding how to handle
>> this situation.
>>
>> How to resolve this issue?
>>
>> The pgxc_ctl in its binary version lacks Dropdb and Dropuser commands?
>> I must choose between the binary version and the version bash? What would
>> be the impact of this change ?
>>
>> Can anyone guide me please
>>
>> Thanks in advance,
>>
>> Hector M. Jacas
>>
>> ---
>> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE
>> running at host imx3.etecsa.cu
>> Visit our web-site: <http://www.kaspersky.com>, <http://www.viruslist.com
>> >
>>
>>
>> ------------------------------------------------------------------------------
>> October Webinars: Code for Performance
>> Free Intel webinars can help you accelerate application performance.
>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
>> from
>> the latest Intel processors and coprocessors. See abstracts and register >
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Postgres-xc-general mailing list
>> Pos...@li...
>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>
>
> ---
> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE running at host imx2.etecsa.cu
>
> Visit our web-site: <http://www.kaspersky.com> <http://www.kaspersky.com>, <http://www.viruslist.com> <http://www.viruslist.com>
>
>
>
> ---
> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE
> running at host imx3.etecsa.cu
> Visit our web-site: <http://www.kaspersky.com>, <http://www.viruslist.com>
>
>


-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company

Please do not worry about it.   It is more than happy to hear any
requirements/good to have things.   Pgxc_ctl is not a complicated product
and you can submit patches.

Regards;

---
Koichi Suzuki


2013/10/10 Hector M. Jacas <hec...@et...>

>  Hi all,
>
> First I must apologize because obviously I was reading the document
> oriented to bash version of pgxc_ctl.
>
> In the documentation of the binary version there is no reference to these
> facilities (dropdb/dropuser). This is the one I should have read .
>
> My mistake and my apologies .
>
> I beg your patience and condescension when from the user role of this
> great project I take the liberty to comment the answer to my post .
>
> I do not share the view of Mr. Ashutosh Bapat when he says " is not an
> interface pgxc_ctl for dropping database or user. " reason: "It's just a
> cluster management utility "
>
> I think there is an inconsistency in that statement because the same
> reason for not including dropdb and dropuser commands are perfectly valid
> createdb and createuser .
>
> You as project developers ( or contribution ) decide the philosophy with
> which your product works and I in my role as user I should be able to
> reconcile my working methods with the philosophy of the tools I have
> selected.
>
> POSTGRESXC is a great project because it solves big problems.
>
> PGXC_CTL is another great project because it simplifies the deployment and
> management of postgresxc and if you add shortcuts to frequently used
> commands (and perhaps, some security features) this project could become a
> kind of Central Command for POSTGRESXC .
>
> Thank you very much for your answers ,
>
> Hector M. Jacas
>
>
>
>
> On 10/09/2013 12:12 AM, Ashutosh Bapat wrote:
>
>  Hector,
>  AFAIK, pgxc_ctl is not an interface for dropping database or user. It's
> just a cluster management utility. You should use corresponding binaries or
> SQL commands for that purpose.
>
>
> On Tue, Oct 8, 2013 at 9:32 PM, Hector M. Jacas <hec...@et...>wrote:
>
>>
>> Hi all,
>>
>> Among the features described in:
>> https://github.com/koichi-szk/PGXC-Tools/blob/master/pgxc_ctl/manual.txtis deleting the databases (Dropdb) and users (Dropuser) and when I try make
>> use of these commands pgxc_ctl answers: command not found
>>
>> PGXC Createdb testdb
>> Selected coord2.
>> PGXC Dropdb testdb
>> sh: Dropdb: command not found
>> PGXC Createuser usertest1
>> Selected coord1.
>> PGXC Dropuser usertest1
>> sh: Dropuser: command not found
>> PGXC
>>
>> Carefully review the source code and found that in the folder:
>> postgres-xc/contrib/pgxc_ctl , there is a file (do_command.c) in which
>> reference is made and performed the execution of Createdb (line 2339) and
>> Createuser (line 2369).
>>
>> In this file there is no reference whatsoever to Dropdb or Dropuser .
>>
>> There is another file (in the same directory) called: pgxc_ctl.bash, in
>> which reference is made and run the corresponding command to Createdb,
>> Dropdb, Createuser and Dropuser.
>>
>> Do not remember reading during pgxc compliacion and deployment (or
>> pgxc_ctl in the area of contributions ) anything regarding how to handle
>> this situation.
>>
>> How to resolve this issue?
>>
>> The pgxc_ctl in its binary version lacks Dropdb and Dropuser commands?
>> I must choose between the binary version and the version bash? What would
>> be the impact of this change ?
>>
>> Can anyone guide me please
>>
>> Thanks in advance,
>>
>> Hector M. Jacas
>>
>> ---
>> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE
>> running at host imx3.etecsa.cu
>> Visit our web-site: <http://www.kaspersky.com>, <http://www.viruslist.com
>> >
>>
>>
>> ------------------------------------------------------------------------------
>> October Webinars: Code for Performance
>> Free Intel webinars can help you accelerate application performance.
>> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
>> from
>> the latest Intel processors and coprocessors. See abstracts and register >
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Postgres-xc-general mailing list
>> Pos...@li...
>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EnterpriseDB Corporation
> The Postgres Database Company
>
>
> ---
> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE running at host imx2.etecsa.cu
>
> Visit our web-site: <http://www.kaspersky.com> <http://www.kaspersky.com>, <http://www.viruslist.com> <http://www.viruslist.com>
>
>
>
> ---
> This message was processed by Kaspersky Mail Gateway 5.6.28/RELEASE
> running at host imx3.etecsa.cu
> Visit our web-site: <http://www.kaspersky.com>, <http://www.viruslist.com>
>
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>
>