sklee2 / solrmarc

Automatically exported from code.google.com/p/solrmarc
0 stars 0 forks source link

import failure #1

Open GoogleCodeExporter opened 8 years ago

GoogleCodeExporter commented 8 years ago
What steps will reproduce the problem?
1. Install SolrMarc in Vufind 1.0 inside directory "indexer" 
2. Solrpath = c:\vufind\solr; url = 
http://localhost:8080/solr/biblio/update
3. Import .mrc file using indexfile command

What is the expected output? What do you see instead?
Output is:
2009-04-29 16:10:10,503 [main] INFO  main org.solrmarc.marc.MarcImporter - 
Exception occurred while Indexing: Error: Problem invoking getDoc in 
SolrCoreProxy
2009-04-29 16:10:10,503 [main] INFO  main org.solrmarc.marc.MarcImporter - 
Setting Solr closed flag
2009-04-29 16:10:10,518 [main] INFO  main org.solrmarc.marc.MarcImporter -
  Adding 0 of 1 documents to index
2009-04-29 16:10:10,518 [main] INFO  main org.solrmarc.marc.MarcImporter -
  Deleting 0 documents from index
2009-04-29 16:10:10,594 [main] INFO  main org.solrmarc.marc.MarcImporter - 
<?xml version="1.0" encoding="UTF-8"?>
2009-04-29 16:10:10,594 [main] INFO  main org.solrmarc.marc.MarcImporter - 
<response>
2009-04-29 16:10:10,594 [main] INFO  main org.solrmarc.marc.MarcImporter - 
<lst name="responseHeader"><int name="status">0</int><int 
name="QTime">15</int></lst>
2009-04-29 16:10:10,610 [main] INFO  main org.solrmarc.marc.MarcImporter - 
</response>
2009-04-29 16:10:10,610 [main] INFO  main org.solrmarc.marc.MarcImporter - 
Finished indexing in 0:00.00
2009-04-29 16:10:10,625 [main] INFO  main org.solrmarc.marc.MarcImporter - 
Indexed 0 at a rate of about 0.0 per sec
2009-04-29 16:10:10,625 [main] INFO  main org.solrmarc.marc.MarcImporter - 
Deleted 0 records
2009-04-29 16:10:10,640 [Thread-1] INFO  Thread-1 
org.solrmarc.marc.MarcImporter - Starting Shutdown hook
2009-04-29 16:10:10,640 [Thread-1] INFO  Thread-1 
org.solrmarc.marc.MarcImporter - Finished Shutdown hook

What version of the product are you using? On what operating system?
Binary_Generic_VuFind_SolrMarc_PC.zip on Windows Server 2003

Please provide any additional information below.

Also tried installing in Vufind 8.0 without success 

Original issue reported on code.google.com by monsonj...@gmail.com on 30 Apr 2009 at 9:55

GoogleCodeExporter commented 8 years ago
Can you provide some more detailed information?   I tried the same sequence of 
steps
that you outlined above, and the indexer seemed to work, producing the 
following results:

 INFO [main] (MarcImporter.java:516) - Starting SolrMarc indexing.
 INFO [main] (SolrCoreLoader.java:98) - Using the data directory of:
C:/Development/vufind/vufind-1.0RC1/solr/biblio
 INFO [main] (SolrCoreLoader.java:100) - Using the multicore schema file at : C:
\Development\vufind\vufind-1.0RC1\solr\solr.xml
 INFO [main] (SolrCoreLoader.java:101) - Using the biblio core
May 4, 2009 3:32:11 PM org.apache.solr.core.SolrCore initDeprecatedSupport
WARNING: solrconfig.xml uses deprecated <admin/gettableFiles>, Please update 
your
config to use the ShowFileRequestHandler.
May 4, 2009 3:32:11 PM org.apache.solr.core.SolrCore initDeprecatedSupport
WARNING: adding ShowFileRequestHandler with hidden files: [.SVN]
May 4, 2009 3:32:11 PM org.apache.solr.core.SolrCore initDeprecatedSupport
WARNING: solrconfig.xml uses deprecated <admin/gettableFiles>, Please update 
your
config to use the ShowFileRequestHandler.
May 4, 2009 3:32:11 PM org.apache.solr.core.SolrCore initDeprecatedSupport
WARNING: adding ShowFileRequestHandler with hidden files: [.SVN]
 INFO [main] (MarcImporter.java:240) - Added record 1 read from file: AAA0000
 INFO [main] (MarcImporter.java:240) - Added record 2 read from file: AAA0001
 INFO [main] (MarcImporter.java:240) - Added record 3 read from file: AAA0002
 INFO [main] (MarcImporter.java:455) -  Adding 3 of 3 documents to index
 INFO [main] (MarcImporter.java:456) -  Deleting 0 documents from index
 INFO [main] (MarcImporter.java:336) - Calling commit
 INFO [main] (MarcImporter.java:347) - Done with the commit, closing Solr
 INFO [main] (MarcImporter.java:350) - Setting Solr closed flag
 WARN [main] (MarcImporter.java:381) - Unable to establish connection to solr server
at URL: http://localhost:8983/solr/update
 INFO [main] (MarcImporter.java:465) - Finished indexing in 0:01.00
 INFO [main] (MarcImporter.java:474) - Indexed 3 at a rate of about 2.0 per sec
 INFO [main] (MarcImporter.java:475) - Deleted 0 records
 INFO [Thread-1] (MarcImporter.java:404) - Starting Shutdown hook
 INFO [Thread-1] (MarcImporter.java:423) - Finished Shutdown hook

Try running the indexer with the attached sample marc record file that contains 
three
records.

Note that it seems that you must run the indexer from the top level vufind 
directory
via the command:    indexer\indexfile three_recs.mrc  otherwise as solr starts 
up
creating its separate solr cores (authority and stats) in a location relative to
where you started it rather than relative to the solr home you specify.  (This 
is not
a problem with solrmarc, rather it seems to be an issue pertaining to solr 
running
with multiple cores)

Original comment by rh...@virginia.edu on 4 May 2009 at 7:46

Attachments:

GoogleCodeExporter commented 8 years ago
Thank you! Your "three_recs.mrc" works, but our MARC records of the Project 
Gutenberg
does not work. We get the record from 
http://www.cucat.org/library/pgmarc.mrc.zip

Does the MARC record have to have a "001" field in order to be imported?

Are you also familiar with VuFind? Why is that we can not find the imported 
records
through the interface?

Thank you!

Original comment by mingluw...@gmail.com on 5 May 2009 at 7:32

GoogleCodeExporter commented 8 years ago
Although MARC records do not require a 001 for them to be valid MARC records, 
Solr
does require a unique identifier for each record that is added to it.   
Furthermore,
the version of SolrMarc that you have uses the VuFind standard for specifying 
how
MARC records are added to solr, which specifies that the unique identifier 
should be
read from the 001 field.  

However, if the value in the 830v field is a unique identifier, with a 
relatively
simple change to the indexing specification, you should be able to index these 
records.

Steps:
1)  cd to indexer directory where you unpacked the SolrMarc distribution.

2)  .\getindexfile

3)  edit vufind_index.properties file  to change this line:
id = 001, first
to this:
id = 830v, first

4) put indexfile back in the jar:
jar uf Generic_VuFind_SolrMarc.jar vufind_index.properties

5) test indexer:
./indextest pgmarc.mrc 

6) cd back to main VuFind Directory

7) run indexer:
indexer\indexfile pgmarc.mrc

8) cross fingers and hope for the best  :-)

-Bob

Original comment by rh...@virginia.edu on 5 May 2009 at 9:09

GoogleCodeExporter commented 8 years ago
Thanks a lot!  We've managed to successfuly index the pgmarc.mrc file.  However 
none 
of the records are showing up in the VuFind interface.  I took a cursory look 
at the 
vufind tech list and didn't see anyone else with this problem.  Have you ever 
seen 
this happen before?

Original comment by monsonj...@gmail.com on 6 May 2009 at 7:45

GoogleCodeExporter commented 8 years ago
I can't help you with any problems with VuFind itself, but I can help you 
determine
whether it is a problem with solrmarc creating the index, or with vufind 
reading the
index.  

Assuming that you are running the VuFind in the standard way, you should be 
able to
access the solr admin page via a web browser.  Goto to the URL:

http://localhost:8983/solr/admin

Or if you can't run a web browser on the machine running VuFind, go to the URL:

http://fullmachine.name.running.vufind:8983/solr/admin

In the large white box labeled Query String:  enter  the following:

[* TO *]

and press the search button.

That should search for all documents in the solr index, if there are no 
documents
there, the problem is in SolrMarc, and I will try to help you further.  If it 
does
show documents, the problem is in VuFind somewhere.

-Bob

Original comment by rh...@virginia.edu on 6 May 2009 at 8:17

GoogleCodeExporter commented 8 years ago
The query returned this:

 <?xml version="1.0" encoding="UTF-8" ?> 
- <response>
- <lst name="responseHeader">
  <int name="status">0</int> 
  <int name="QTime">15</int> 
  </lst>
  <result name="response" numFound="0" start="0" /> 
  </response>

Original comment by monsonj...@gmail.com on 6 May 2009 at 9:20

GoogleCodeExporter commented 8 years ago
Hmmm.   That certainly looks like a problem on the indexing end.  Although it 
may be
that the indexer is writing its index in one place, whereas the jetty is 
looking in
another place for the index.   I'll see if I can figure it out tomorrow.  
Although
I'll probably need to try to get VuFind running on my windows dev box, which 
doesn't
seem exactly straightforward.

One thing that might be worth trying is with the solr server running, execute 
the
command:

indexer\setsolrhome  http://localhost:8983/solr

which will talk to the solr server directly to get the correct value for 
solr.home

then try running the indexer again:

indexer\indexfile pgmarc.mrc

and then try the query in solr again.

-Bob

Original comment by rh...@virginia.edu on 6 May 2009 at 9:50

GoogleCodeExporter commented 8 years ago
I have something else for you to try.  The above suggestion won't work for you.
THere is a problem with the set solr home implementation when dealing with a
multicore Solr index such as is used by VuFind.

Instead try this:

indexer\setsolrhome  http://localhost:8983/solr/biblio ./solr

after which the indexer should work:

indexer\indexfile pgmarc.mrc

If I'm right the problem wasn't the the records weren't being added to the 
correct
location, the problem was that the solr search engine running in jetty wasn't 
being
told that there were new records in the index.

Original comment by rh...@virginia.edu on 8 May 2009 at 1:46

GoogleCodeExporter commented 8 years ago
Bob,

We tried the methods you mentioned, and I did see a new "index" folder with 
files
created after indexing the marc file. But still no records could be found in the
solr/admin

So, is that because "jetty wasn't being told that there were new records in the
index"? How to fix this?

Original comment by mingluw...@gmail.com on 9 May 2009 at 2:12

GoogleCodeExporter commented 8 years ago
In the following command:

indexer\setsolrhome  http://localhost:8983/solr/biblio ./solr

the first parameter tells where the solr server is currently running in jetty or
tomcat (or some other java webapp container)  the second tells where the solr 
home is
relative to the current directory.  The program uses this first parameter 
internally
to send a  commit  message to the running solr server, which tells it to 
re-read its
index files, taking note of any newly added records.

Where is the new "index" folder that you see after indexing?   It should be in 
the
directory  ..../vufind/solr/biblio  directory    assuming that  ..../vufind  is 
the
top level of the vufind installation.   

Original comment by rh...@virginia.edu on 12 May 2009 at 4:47