[Pgpool-general] Mismatch among backends

[Jaume Sabater]

On Sat, Mar 7, 2009 at 12:41 AM, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:

> In this case you should have an ERROR message in your slave PostgreSQL
> log. That would help what was happening. I bet the table
> TSearcherServices does not exist or the file
> /opt/pgpool2/TSearcherServices.csv.

You were right, Tatsuo. The file that was being used for the COPY
operation was not available in the second node. Quite straightforward
and silly from my side, now that I see what happened. Here you are the
PostgreSQL log:

2009-03-06 08:10:27 CET ERROR:  could not open file
"/opt/pgpool2/TSearcherServices.csv" for reading: No such file or
directory
2009-03-06 08:10:27 CET STATEMENT:  COPY "TSearcherServices"
("IdSearcherServices", "IdSearcher" ,"IdService", "SearcherNumber" )
2009-03-06 08:10:27 CET LOG:  unexpected EOF on client connection
2009-03-06 08:10:27 CET LOG:  unexpected EOF on client connection
2009-03-06 08:10:27 CET LOG:  unexpected EOF on client connection
2009-03-06 08:10:27 CET LOG:  unexpected EOF on client connection


For some stupid reason I had thought that pgpool-II would pick up the
file and pass the content to the nodes, while what happened was that
it just forwarded the sentence to each node for them to execute it
(which makes sense, don't get me wrong: my fault). :)

-- 
Jaume Sabater
http://linuxsilo.net/

"Ubi sapientas ibi libertas"