postgres-xc-developers Mailing List for Postgres-XC

I applied only that fix for now since these are two different bugs

Sent from my iPhone

On 03-Oct-2011, at 6:41 PM, Michael Paquier <mic...@gm...> wrote:

> Has the patch been correctly applied on master?
> This commit mail, gitweb, and the commit fetched look kind of weird, only the diffs in explain.c appear.
> -- 
> Michael Paquier
> http://michael.otacoo.com
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure contains a
> definitive record of customers, application performance, security
> threats, fraudulent activity and more. Splunk takes this data and makes
> sense of it. Business sense. IT sense. Common sense.
> http://p.sf.net/sfu/splunk-d2dcopy1
> _______________________________________________
> Postgres-xc-developers mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers

Has the patch been correctly applied on master?
This commit mail, gitweb, and the commit fetched look kind of weird, only
the diffs in explain.c appear.
-- 
Michael Paquier
http://michael.otacoo.com

Hi Michael,
I was able to run the regression (make installcheck with 2 datanodes, 1
coordinator) 4 times in row without rebuilding data cluster or re-booting
the XC instances. So, it looks like the patch you gave has fixed the
assertion failure, which I could observe in two successive runs of
regression.

On Mon, Oct 3, 2011 at 11:30 AM, Ashutosh Bapat <
ash...@en...> wrote:

> I am running regression with your previous patch. It consistently gave me
> Failed to get pooled connections error. So, I rebuilt data cluster and now
> the regression is running. I will try again to see if the issue is gone.
>
> On Mon, Oct 3, 2011 at 11:26 AM, Michael Paquier <
> mic...@gm...> wrote:
>
>> Hi,
>>
>> Please find attached the patch fixing all the regression issues and stable
>> with DBT-1.
>> I think this patch is good for commit.
>>
>> Regards,
>>
>> --
>> Michael Paquier
>> http://michael.otacoo.com
>>
>>
>> ------------------------------------------------------------------------------
>> All the data continuously generated in your IT infrastructure contains a
>> definitive record of customers, application performance, security
>> threats, fraudulent activity and more. Splunk takes this data and makes
>> sense of it. Business sense. IT sense. Common sense.
>> http://p.sf.net/sfu/splunk-d2dcopy1
>> _______________________________________________
>> Postgres-xc-developers mailing list
>> Pos...@li...
>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>>
>>
>
>
> --
> Best Wishes,
> Ashutosh Bapat
> EntepriseDB Corporation
> The Enterprise Postgres Company
>
>


-- 
Best Wishes,
Ashutosh Bapat
EntepriseDB Corporation
The Enterprise Postgres Company

I am running regression with your previous patch. It consistently gave me
Failed to get pooled connections error. So, I rebuilt data cluster and now
the regression is running. I will try again to see if the issue is gone.

On Mon, Oct 3, 2011 at 11:26 AM, Michael Paquier
<mic...@gm...>wrote:

> Hi,
>
> Please find attached the patch fixing all the regression issues and stable
> with DBT-1.
> I think this patch is good for commit.
>
> Regards,
>
> --
> Michael Paquier
> http://michael.otacoo.com
>
>
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure contains a
> definitive record of customers, application performance, security
> threats, fraudulent activity and more. Splunk takes this data and makes
> sense of it. Business sense. IT sense. Common sense.
> http://p.sf.net/sfu/splunk-d2dcopy1
> _______________________________________________
> Postgres-xc-developers mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-developers
>
>


-- 
Best Wishes,
Ashutosh Bapat
EntepriseDB Corporation
The Enterprise Postgres Company

Hi,

Please find attached the patch fixing all the regression issues and stable
with DBT-1.
I think this patch is good for commit.

Regards,
-- 
Michael Paquier
http://michael.otacoo.com

Hi,

Here is a patch stabilizing the regressions for the issue with curval.
The issue disappears when running multiple regression tests, but there is
still an issue with the regression test privileges due to permissions on the
database when creating a schema. I am digging into that but it looks to be
an old issue.
Could you check that your regressions got better?
-- 
Michael Paquier
http://michael.otacoo.com

Hi,

I found that the problem is a little bit more generic.
When you disconnect a backend on a remote XC node, with a CLEAN CONNECTION,
you need to close a pooler connection.
I found that a PGXC node always cleans up its session if a temporary object
has been used during the session.
Cleaning up means the usage of DISCARD queries automatically, so it means
that a transaction ID has to be consumed.
I am changing that in a way such as when a closing session on a remote XC
node looks for a transaction ID from GTM and closes nicely.
-- 
Michael Paquier
http://michael.otacoo.com