[Koha] Import staged MARC records with ElasticSearch engine

[Mark Alexander]

Excerpts from Abdulsalam Yousef's message of 2017-04-17 11:04:58 +0300:
> - Importing MARC records into catalog form (Manage staged MARC records) in
> staff client GUI using ElasticSearch engine is very slow and accidentally
> stops after importing some hundreds of records, That we may spend 3 days to
> import about 26000 record only, Is that normal? or there are any other
> configurations should be done first with ElasticSearch engine?
> 
> - We are using Koha 16.11.01.

We're not using Elastic Search on our installation, but I have found
that attempting to import our entire catalog of 24000 MARC records
failed due to Koha running out of memory (our system has 2 GB of RAM).
I was forced to split the MARC file into smaller chunks of 2000
records, and import each chunk separately.  A larger chunk might have
worked but I didn't want to take any chances.

I wrote a Ruby script to do the splitting of the MARC records, but
perhaps other tools can do the job (maybe Marcedit, which I haven't used).
The Ruby script is here:

https://github.com/bloovis/marc/blob/master/splitmarc.rb