Re: [Postgres-xc-developers] Fix for xc_FQS_join regression

[Koichi S. <koi...@gm...>]

Thanks a lot;

----------
Koichi Suzuki


2013/5/15 Ashutosh Bapat <ash...@en...>

> I committed this patch, with slight changes. I tested it on build farm to
> make sure that the test passes there as well.
>
>
> On Wed, May 15, 2013 at 2:07 PM, 鈴木 幸市 <ko...@in...> wrote:
>
>> Sorry for bothering.   This time, I only applied  num_nodes off only to
>> the statement which is shipped to more than one datanode.  There was two of
>> them.
>>
>> Hope this makes sense.
>>
>> Regards;
>>  ---
>> Koichi Suzuki
>>
>>
>>
>> On 2013/05/15, at 15:04, Ashutosh Bapat <ash...@en...>
>> wrote:
>>
>> Hi Suzuki-san,
>>
>>
>>
>> On Wed, May 15, 2013 at 11:15 AM, 鈴木 幸市 <ko...@in...> wrote:
>>
>>> I see.   PFA the revised patch.   Some num_node off was missing in
>>> explain statements.  They're fixed too.
>>>
>>>
>>>
>> You need to apply changes only to the queries which are showing diffs
>> because of difference in number of nodes. But it seems you have changed all
>> the explain commands in the file. That's not needed.
>>
>> As I said in my previous mail, there are some explain outputs where the
>> query is shipped to only one node and the node count in explain output will
>> be always 1 for such queries. That doesn't change with the cluster
>> configuration. For such queries, it's important that we have num_nodes off.
>>
>>
>>>  ---
>>> Koichi Suzuki
>>>
>>>
>>>
>>> On 2013/05/15, at 13:17, Ashutosh Bapat <ash...@en...>
>>> wrote:
>>>
>>> That's not a correct fix. We should turn off the node count in Explain
>>> output. Use EXPLAIN  (verbose, num_nodes off, nodes off, costs off) as a
>>> general rule. In case of single node reduction tests, we may use num_nodes
>>> on, but in that case num_nodes will always be 1, and will not change
>>> depending upon cluster configuration.
>>>
>>>
>>> On Wed, May 15, 2013 at 7:56 AM, Koichi Suzuki <
>>> koi...@gm...> wrote:
>>>
>>>> I found the regression failure was due to the change of explain output.
>>>>   Original expected output uses node count = 3, which is not correct
>>>> because regression test is configured with only two datanode.   Actual
>>>> result says node count = 2 which is correct.
>>>>
>>>> Attached is a patch for the correction.
>>>>
>>>> Regards;
>>>> ----------
>>>> Koichi Suzuki
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> AlienVault Unified Security Management (USM) platform delivers complete
>>>> security visibility with the essential security capabilities. Easily and
>>>> efficiently configure, manage, and operate all of your security controls
>>>> from a single console and one unified framework. Download a free trial.
>>>> http://p.sf.net/sfu/alienvault_d2d
>>>> _______________________________________________
>>>> Postgres-xc-developers mailing list
>>>> Pos...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>>>>
>>>>
>>>
>>>
>>> --
>>> Best Wishes,
>>> Ashutosh Bapat
>>> EntepriseDB Corporation
>>> The Postgres Database Company
>>> ------------------------------------------------------------------------------
>>> AlienVault Unified Security Management (USM) platform delivers complete
>>> security visibility with the essential security capabilities. Easily and
>>> efficiently configure, manage, and operate all of your security controls
>>> from a single console and one unified framework. Download a free trial.
>>>
>>> http://p.sf.net/sfu/alienvault_d2d_______________________________________________
>>> Postgres-xc-developers mailing list
>>> Pos...@li...
>>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>>>
>>>
>>>
>>>
>>
>>
>> --
>> Best Wishes,
>> Ashutosh Bapat
>> EntepriseDB Corporation
>> The Postgres Database Company
>>
>>
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EntepriseDB Corporation
> The Postgres Database Company
>