Hi all,

While trying to tune Zebra for a large Koha library, I ran into some problems when throwing memory at the memMax configuration option.

While the documentation says it can use up to half the memory on your system, I’ve discovered the practical limit is actually 2047MB.

Why 2047MB? It’s because memMax gets puts into a variable with an “int” data type which is a signed 32 bit integer. If you get to 2048MB or over, you’ll end up with invalid values that cause memory allocation issues and your indexing will fail.

I’ve raised an issue with Indexdata to either switch to an unsigned 32 bit integer, switch to a 64 bit integer, or at least add some error checking: https://github.com/indexdata/idzebra/issues/34

But I figured I’d share my little discovery with y’all in the meantime.

David Cook

Senior Software Engineer

Prosentient Systems

Suite 7.03

6a Glen St

Milsons Point NSW 2061

Australia

Office: 02 9212 0899

Online: 02 8005 0595