Re: [Postgres-xc-general] "Tuple not found error" during Index creation - Mailing list pgsql-general

From Mason Sharp
Subject Re: [Postgres-xc-general] "Tuple not found error" during Index creation
Date
Msg-id CA+A7BJF1-u9ZJjn9GUCAnpeb8xq-XLJKW5QLR+kX0M2x1=N_9A@mail.gmail.com
Whole thread Raw
In response to Re: [Postgres-xc-general] "Tuple not found error" during Index creation  (Michael Paquier <michael.paquier@gmail.com>)
List pgsql-general



On Tue, Dec 10, 2013 at 11:23 PM, Michael Paquier <michael.paquier@gmail.com> wrote:
On Wed, Dec 11, 2013 at 1:17 PM, Sandeep Gupta <gupta.sandeep@gmail.com> wrote:
> Hi Michael,
>
>  I can provide the table schema and the data over which indexing almost
> always fails with tuple not found error.
> Would this be of help.  The other issue is that file is 3.2GB  so we would
> have work some logistics to transfer this across.
Transferring a data file of a couple of gigs is out of question. My
point was to know if you are able to create a self-contained test case
using automatically generated data of this type:
create table foo as select generate_series(1,100000000) as a, 'bbbb'::text as b;
create index fooi on aa(a);
Or smth like that.
This way you wouldn't need to 1) publish your schema, 2) transfer huge
files of data. And this would make tracking of this error somewhat
easier.

Perhaps I could help Sandeep. Sandeep, before working out transferring such a large file though, please instead try to pinpoint the particular commit that introduced the issue.

I could recently reproduce a problem with VACUUM FULL in XC 1.1 (stop cluster, restart cluster, execute VACUUM FULL).  I could not reproduce the problem however with a build from the commit 11339220012a9e73cb82039b0ad41afd71bafca2 on Aug 22, 2012.  I suspect that a commit after that one to procarray.c may have something to do with the problem. It may be a similar issue for you. If you have the time, perhaps you can try different commits based on the git log of procarray.c to see if any particular one introduced the issue. 

I would use the commit not just on procarray, but on the entire source tree. If you find a suspect one based on procarray.c, try the one immediately preceding it for the whole source tree to confirm that it is procarray-related.  

Again, this all may be unrelated, but I think it is a good place to start, so I apologize in advance if this chews up some of your time unnecessarily.  
 
Regards,
--
Michael

------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT
organizations don't have a clear picture of how application performance
affects their revenue. With AppDynamics, you get 100% visibility into your
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Postgres-xc-general mailing list
Postgres-xc-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general



--
Mason Sharp

TransLattice - http://www.translattice.com
Distributed and Clustered Database Solutions


pgsql-general by date:

Previous
From: Koichi Suzuki
Date:
Subject: Re: [Postgres-xc-general] "Tuple not found error" during Index creation
Next
From: John Melesky
Date:
Subject: Re: Zero dead tuples, when significant apparent bloat