On Mon, Jun 10, 2013 at 6:19 PM, Yamil Suarez <[email protected]> wrote: > Hello, > > To maintaining test VMs with production data I usually use a combination of > pg_dump[1] on the production server. I then I use "dropdb evergreen" and then > pg_restore[2] on the test VMs. On occasion I get an error message of > "segmentation fault" when I try to run pg_restore[2] on the test VM, but > usually I don't have any issues. Before I got this error today, I did a > pg_dump of the concerto data that was already on the VMs postgres, and as a > test I successfully used pg_restore to reload the concerto data with out any > problems. Except for some warning I usually get when I successfully run > pg_restore[3]. Finally, I checked /var/log/postgresql/postgresql-9.1-main.log > and only saw entries for the warnings I just mentioned from when I restored > up the concerto data as a test. > > I was wondering if someone had an idea of why this is happening to me this > morning? > > Thanks in advance, > Yamil > > > [1] pg_dump -Fc evergreen > prod_data.sql > > > [2] pg_restore -U postgres -C -j 4 --dbname=template1 prod_data.sql > > > [3] > WARNING: => is deprecated as an operator name > This name may be disallowed altogether in future versions of PostgreSQL.
For the list record, I figured out that I was getting the error because when I used a certain GUI FTP tool and then saved the file on an OS X machine, the pg_dump file was being changed in some way. I confirmed this by checking the MD5 of the file. When I copied the file in a different way, pg_restore did not give me an error anymore. Yamil
