[SOLR-11693] Class loading problem for Tika/POI for some PPTX

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 7.1
• Fix Version/s: 7.3
• Component/s: contrib - DataImportHandler
• Assignee: Unassigned
• Reporter: Tim Allison
• Created: 28/Nov/17 21:18
• Updated: 08/Jun/19 15:16
• Resolved: 25/Dec/17 01:45

Description

advokat reported TIKA-2497. I can reproduce this issue with a Solr instance in both 6.6.2 and 7.1.0.

I can't reproduce it when I run the triggering file within Solr's unit tests or with straight Tika.

Would anyone with more knowledge of classloading within Solr be able to help?

See TIKA-2497 for triggering file and conf files.

...turns out this is a bug in POI 3.16 and 3.17-beta1

Issue Links

• depends upon: SOLR-11701 Upgrade to Tika 1.17 when available
• relates to: TIKA-2497 Unexpected RuntimeException from org.apache.tika.parser.microsoft.ooxml.OOXMLParser
• relates to: SOLR-11622 Bundled mime4j library not sufficient for Tika requirement