[Koha] Staging of import stalls
kmkale at anantcorp.com
kmkale at anantcorp.com
Sun Mar 1 20:02:26 NZDT 2009
Ed,
I also had issues getting zebra working initially as we had most of the
biblios in Unicode Marathi.So while I was playing with zebra we used
nozebra to get the services off the ground and then switched to zebra once
we had solved the zebra search of Unicode Marathi.
1) Turn off noZebra from system preferences.
For testing Zebra you can run it from command line.
sudo -u ${KOHA_USER} zebrasrv -f /etc/koha/koha-conf.xml
adjust the path to koha-conf.xml if different.
The first time I ran this it refused to run as it was not finding some
directories. I had to create those and change ownerships to koha user. I
forget exact paths but a look at the output will tell you.
Once its running, do your bulk import by staging and then importing. After
that in a separate terminal run misc/migration_tools/rebuild_zebra.pl -b -w
This should build your index and then you should be able to search from
opac & stff client.
2) Once you have the search going you can run the zebra server and
zebraqueue in daemon mode and have them start on system boot. Here is how
to do it on Ubuntu, adjust according to your OS.
sudo ln -s ${SCRIPT_DIR}/koha-zebra-ctl.sh /etc/init.d/koha-zebra-daemon
${SCRIPT_DIR} is /usr/share/koha/bin/ by default in a standard install.
sudo update-rc.d koha-zebra-daemon defaults
sudo /etc/init.d/koha-zebra-daemon start
Zebraqueue Daemon:
sudo ln -s ${SCRIPT_DIR}/koha-zebraqueue-ctl.sh
/etc/init.d/koha-zebraqueue-daemon
sudo update-rc.d koha-zebraqueue-daemon defaults
sudo /etc/init.d/koha-zebraqueue-daemon start
Thats it. the zebraqueue will take care of indexing any subsequent
additions you make. If you bulk import any more biblios run
rebiuld_zebra.pl again as zebraqueue is slower in indexing. Also run
rebuild_zebra.pl for your authorities.
The daemons will log in /var/log/koha/ Have a look at files in there if
something doesn't work out.
--
Regards,
Koustubha Kale
Anant Corporation
Contact Details :
Address : 103, Armaan Residency, R. W Sawant Road, Nr. Golden Dyes Naka,
Thane (w),
Maharashtra, India, Pin : 400601.
TeleFax : +91-22-21720108, +91-22-21720109
Mobile : +919820715876
Website : http://www.anantcorp.com
Original Message:
-----------------
From: Roche III, Edward edward_roche at solanco.k12.pa.us
Date: Sat, 28 Feb 2009 22:33:19 -0500
To: kmkale at anantcorp.com, dearden at sarsf.org, koha at lists.katipo.co.nz
Subject: RE: [Koha] Staging of import stalls
how would you get the zebra server running as well as the daemans? I
originally tries to use zebra but it was giving me issues so I decide to go
the nozebra route.
Thanks
Ed
Edward Roche
Technology Support/Network Admin
Solanco School District
edward_roche at solanco.k12.pa.us
717.786.2151
The greatest mistake you can make in life is to continually be afraid you
will make one.
Elbert Hubbard
-----Original Message-----
From: Koustubha Kale [mailto:kmkale at anantcorp.com]
Sent: Sat 2/28/2009 2:31 AM
To: Doug Dearden
Cc: Roche III, Edward; koha at lists.katipo.co.nz
Subject: Re: [Koha] Staging of import stalls
ED,
Do try Doug's suggestion.
I have also faced this problem. removing the problem record solves the
stall.
Also I have staged and imported more than 87000 records in one go, so
the batch process does work well.
In case you are not using zebra, one trick i observed was to have the
zebra server and queue daemons running during the import and put
nozebra to off even if you are not using zebra. this speeds up the
staging and import by orders of magnitude. after the import is over you
can run the rebuild_nozebra.pl to rebuild the index after turning on
nozebra setting. This little trick resulted in hours of time saving.
--
Regards,
Koustubha Kale
Anant Corporation
Contact Details :
Address : 103, Armaan Residency, R. W Sawant Road, Nr. Golden Dyes Naka,
Thane (w),
Maharashtra, India, Pin : 400601.
TeleFax : +91-22-21720108, +91-22-21720109
Mobile : +919820715876
Website : http://www.anantcorp.com
Doug Dearden wrote:
> Hi Ed,
>
> I had the same problem. Take a look at the last import (Home > Tools >
Manage Staged MARC Records > Batch
>
> Figure out the last record in the batch that went in, then edit the
import file you are using and remove the record immediately after that one
in the file. It has some kind of problem. Try again and see if you get
further. It is either that record, or the last one that actually worked, I
can't remember exactly which.
>
> Also, there is a command line tool that you can use to import. You will
find it at /usr/share/koha/bin/migration_tools/bulkmarcimport.pl
>
> >From the migration tools subdirectory enter: perl bulkmarcimport.pl and
you will get the syntax and parameter info. It has a test mode you can run
against the file.
>
> Best,
>
> Doug
>
>
>>>> "Roche III, Edward" <edward_roche at solanco.k12.pa.us> 2/27/2009 9:20 AM
>>>
>>>>
> Good Morning All
>
>
>
> I am trying stage my MARC import data but it keeps stalling. I broke it
> down into 4 individual files of about 4000 records each. But it still
> stalls at about 15%. Has anyone had this issue? Can it be done from the
> terminal instead of through the browser? Every time I start the process
> it pins my cpu at 95 to 100% which is why it is probably stalling but
> not sure why.
>
>
>
> My Config
>
> ----------------------------------
>
> Ubuntu 8.1
>
> 1.5gb memory
>
> Pentium 4 - 2.5ghz dual core
>
> Koha 3.0
>
> NoZebra
>
> Basic LAMP install
>
>
>
> Any thoughts would be great.
>
>
>
> Thanks
>
> Ed
>
>
>
> Edward J. Roche
>
> Network Administrator
>
> Solanco School District
>
> 717.786.2151 x2437
>
>
>
>
>
> "At the End of the Day, Did You Play to Win or Not to Lose?"
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Koha mailing list
> Koha at lists.katipo.co.nz
> http://lists.katipo.co.nz/mailman/listinfo/koha
>
>
>
--------------------------------------------------------------------
mail2web LIVE – Free email based on Microsoft® Exchange technology -
http://link.mail2web.com/LIVE
More information about the Koha
mailing list