0

It seems that Apacke Tika 1.24.1 is creating lots of /tmp/MediaDataBox ISO files, and my /tmp partition gets filled up.

What is MediaDataBox ISO file used for?

Can we somehow tell Tika to save it in another directory?

Tika runs in server mode as follows:

java -Xmx3G -jar tika-server.jar -spawnChild --host=hostname.domain.com

freeAR
  • 943
  • 3
  • 18
  • 32
  • How are you calling Apache Tika? Without the code / command line used, it's hard to tell you how to tweak the temp file handling – Gagravarr Oct 05 '20 at 18:38
  • @Gagravarr : Thanks. I've updated my question. – freeAR Oct 05 '20 at 19:34
  • I have a nasty feeling that the Tika Server doesn't have options to control the temp file strategy... You might be best asking on [the Apache Tika users list](https://lists.apache.org/list.html?user@tika.apache.org) for this – Gagravarr Oct 06 '20 at 11:37
  • Tika will sometimes spool a stream to a file when the underlying parser library needs it, but generally should tidy up, so it might be a bug or a missing flag. Best ask on the user list, not all Apache Tika experts check SO (and I think the relevant 2 people don't!) – Gagravarr Oct 06 '20 at 21:06

1 Answers1

0

This example shows how to save temporary files in an alternate directory:

java -Djava.io.tmpdir=/somewhere/tmp -jar tika-server.jar -spawnChild -JXmx3G -JDjava.io.tmpdir=/somewhere/tmp --host=hostname.domain.com

I found useful information in Tika Server docs

freeAR
  • 943
  • 3
  • 18
  • 32