[SOLR-12423] Upgrade to Tika 1.19.1 when available

• Type: Task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 7.6, 8.0
• Assignee: Erick Erickson
• Reporter: Tim Allison
• Created: 30/May/18 01:30
• Updated: 08/Jun/19 15:15
• Resolved: 17/Oct/18 17:26

Description

In Tika 1.19, there will be the ability to call the ForkParser and specify a directory of jars from which to load the classes for the Parser in the child processes. This will allow us to remove all of the parser dependencies from Solr. We’ll still need tika-core, of course, but we could drop tika-app.jar in the child process’ bin directory and be done with the upgrade... no more fiddly dependency upgrades and threat of jar hell.

The ForkParser also protects against ooms, infinite loops and jvm crashes. W00t!

This issue covers the basic upgrading to 1.19.1. For the migration to the ForkParser, see SOLR-11721.

Issue Links

• depends upon: TIKA-2653 Allow users to specify a directory of jars for classloading in ForkParser
• depends upon: TIKA-2655 Allow the RecursiveParserWrapper to work with the ForkParser
• relates to: SOLR-11721 Isolate most of Tika and dependencies into separate jvm
• links to: GitHub Pull Request #468