[TIKA-2780] Intermittent failures in batch mode when STDIN = /tmp/null

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.19.1
• Fix Version/s: 2.0.0, 1.20
• Component/s: batch
• Assignee: Tim Allison
• Reporter: Jeroen
• Created: 10/Nov/18 14:09
• Updated: 13/Nov/18 22:36
• Resolved: 13/Nov/18 19:20

Description

We found this strange behavior in the R wrapper for tika:

We call rtika in batch mode using the -fileList parameter. However if in the java tika process STDIN is set to /dev/null with read permission (O_RDONLY) as is usually the default then, then tika will randomly fail occasionally with this error:

{{  BatchProcess:ParallelFileProcessingResult

{considered=7, added=7, consumed=2, numberHandledExceptions=0, secondsElapsed=1.539, exitStatus=0, causeForTermination='USER_INTERRUPTION'}

}}

As a workaround we now set STDIN to something that is unreadable and then everything works, but this is really strange behavior.