[Koha] zebra search failure and extended UTF-8 characters showing up as ?s and diamonds in opac

[George Masters]

We are having two different issues with our Koha system even after updating to 
the latest version 3.8.1:

1. Zebra searches fail. When turning off zebra we get an "as-marc" method error. 
Ideally, I would like to use zebra so that we can allow our MARC records to be 
accessed by other libraries.

2. We have materials in Greek, Russian and Spanish in our collection and the 
extended UTF-8 characters come up as ?s and black diamonds. To see what I mean 
do a search in our opac for Alexander Schmemann. It is located at 
opac.orthdoxoeducationonline.net

Although, the SAX parser test says it looks good, the 0x value is not the same 
one that Koha wants even though I did move the lines around in the config file.

We are running the server on Ubuntu Linux 12.04 LTS. We have turned on ICU 
indexing on zebra and set the parameters for ICU in the administrative module of 
Koha under Searching. The export lines for PER5LIB and KOHA_CONF appear to be in 
the correct places. Where I think the issue is is that there are multiple areas 
where these parameters must be inserted. Ideally, a system of this complexity 
should only have one area for configuration changes. My concern is that with so 
many PERL modules there are all too many places for Koha to break.

Attempting to run the rebuild index script rebuild_zebra -b -v -r results in an 
error on line 6.

By turning off the zebra server, the opac partially processes searches. My 
suspicion lies in the handling of a MARC record I recently added from the 
Library of Congress Z39.50 server for a book entitled, "Man of God" published by 
the Nicodemus Press of the Russian Orthodox Church outside of Russia in 1994. 
The MARC record may be missing a required field. We are investigating that 
currently. If someone finds out the issue, please let us know.

George Masters
Director
Orthodox Christian Library System