[OPEN-ILS-DEV] Re: pg_loader getting killed

[John Fink]

Okay, it's finished up, but I've got a bit of a problem -- with
pg_loader.pl, I get one sql file, but after I've run the parallel pg_loader,
I get many many many.

I'm following the Gutenberg MARC conversion instructions:

perl parallel_pg_loader.pl -or bre -or mrd -or mfr -or mtfe -or mafe -or
msfe -or mkfe -or msefe -a mrd -a mfr -a mtfe -a mafe -a msfe -a mkfe -a
msefe < full-bibs.ingest > full-bibs.sql

I end up with these files, instead of one.

pg_loader-output.bre.sql   pg_loader-output.mkfe.sql   pg_loader-
output.msfe.sql
pg_loader-output.mafe.sql  pg_loader-output.mrd.sql    pg_loader-
output.mtfe.sql
pg_loader-output.mfr.sql   pg_loader-output.msefe.sql  pg_loader-output.sql

So which one should I load?  Or am I supposed to load all of them?  If so,
does it matter what order I load them in?

jf

On Wed, Feb 20, 2008 at 11:24 AM, John Fink <john.fink at gmail.com> wrote:

> Oh man, I love you guys.  Thanks!  I'll give it a spin and report.
>
> jf
>
>
> On Feb 20, 2008 11:15 AM, Mike Rylander <mrylander at gmail.com> wrote:
>
> > On Feb 20, 2008 10:26 AM, John Fink <john.fink at gmail.com> wrote:
> > > Oh, and this is Evergreen 1.2.1.0, sorry for not stating that earlier.
> > >
> >
> > Grab parallel_pg_loader.pl from trunk* and try that.  It avoids
> > loading the entire file into memory before writing the files, and was
> > built for just such a situation.  It works fine with 1.2.x
> >
> > *
> > http://svn.open-ils.org/trac/ILS/browser/trunk/Open-ILS/src/extras/import/parallel_pg_loader.pl
> >
> > --miker
> >
> > > jf
> > >
> > >
> > >
> > > On Feb 20, 2008 9:18 AM, John Fink <john.fink at gmail.com> wrote:
> > >
> > > > Hey folks,
> > > >
> > > > I've tried a couple of times unsuccessfully to run pg_loader on what
> > is
> > > admittedly a vastly huge (ls shows me filesize 17077683247) ingest
> > file.
> > > For a while it hums merrily along, but then I inevitably get something
> > like
> > > the following:
> > > >
> > > > 8643000 363.398018751032Killed
> > > >
> > > > Complete with a zero-length sql.  Am I simply biting off more than I
> > can
> > > chew?  Is it possible to easily split the ingest file for easier
> > loading?
> > > It looks like to me that pg_loader is trying to load everything into
> > memory
> > > before writing the sql (hence the zero length).
> > > >
> > > > jf
> > > >
> > > >
> > > >
> > > > --
> > > > http://libgrunt.blogspot.com -- library culture and technology.
> > >
> > >
> > >
> > > --
> > > http://libgrunt.blogspot.com -- library culture and technology.
> >
> >
> >
> > --
> > Mike Rylander
> >  | VP, Research and Design
> >  | Equinox Software, Inc. / The Evergreen Experts
> >  | phone:  1-877-OPEN-ILS (673-6457)
> >  | email:  miker at esilibrary.com
> >  | web:  http://www.esilibrary.com
> >
>
>
>
> --
> http://libgrunt.blogspot.com -- library culture and technology.
>



-- 
http://libgrunt.blogspot.com -- library culture and technology.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.georgialibraries.org/pipermail/open-ils-dev/attachments/20080221/cda0f275/attachment.html