Fatal error - memory exhausted

Possibly related to the following issues on older branches:

#2474849: Ignore attachments based on uri, filesize or file extension →
🐛 Request Entity Too Large Closed: cannot reproduce

Perhaps the way forward would be to limit extraction by file size.

There is a config option in the schema, and this is configurable via the Search API processors form.

We saw the error with the following config:

  file_attachments:
    excluded_extensions: 'aif art avi bmp gif ico mov oga ogv png psd ra ram rgb flv'
    number_indexed: 0
    number_first_bytes: '1 MB'
    max_filesize: '0'
    excluded_private: 1
    excluded_mimes: 'audio/x-aiff image/x-jg video/x-msvideo image/x-ms-bmp image/gif image/vnd.microsoft.icon video/quicktime audio/ogg video/ogg image/png image/x-photoshop audio/x-realaudio audio/x-pn-realaudio image/x-rgb video/x-flv'
 

Perhaps we need to try changing these settings?

It may also be worth adding extra logging, similar to 📌 Introduce debug mode - do not pollute indexing output with Tika warnings Active .

For example, if debug mode is on, the extractor could:

1. log details of the file before it tries to do the extraction
2. report success or failure