[TIKA-1179] A corrupt mp3 file can cause an infinite loop in Mp3Parser

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Cannot Reproduce
• Affects Version/s: 1.4
• Fix Version/s: 1.5
• Component/s: parser
• Assignee: Ray Gauss II
• Reporter: Marius Dumitru Florea
• Created: 04/Oct/13 09:01
• Updated: 25/Mar/14 16:21
• Resolved: 04/Oct/13 19:22

Description

I have a thread that indexes (among other things) files using Apache Sorl. This thread hangs (still running but with no progress) when trying to extract meta data from the mp3 file attached to this issue. Here are a couple of thread dumps taken at various moments:

"XWiki Solr index thread" daemon prio=10 tid=0x0000000003b72800 nid=0x64b5 runnable [0x00007f46f4617000]
   java.lang.Thread.State: RUNNABLE
	at org.apache.commons.io.input.AutoCloseInputStream.close(AutoCloseInputStream.java:63)
	at org.apache.commons.io.input.AutoCloseInputStream.afterRead(AutoCloseInputStream.java:77)
	at org.apache.commons.io.input.ProxyInputStream.read(ProxyInputStream.java:99)
	at java.io.BufferedInputStream.fill(Unknown Source)
	at java.io.BufferedInputStream.read1(Unknown Source)
	at java.io.BufferedInputStream.read(Unknown Source)
	- locked <0x00000000cb7094e8> (a java.io.BufferedInputStream)
	at org.apache.tika.io.ProxyInputStream.read(ProxyInputStream.java:99)
	at java.io.FilterInputStream.read(Unknown Source)
	at org.apache.tika.io.TailStream.read(TailStream.java:117)
	at org.apache.tika.io.TailStream.skip(TailStream.java:140)
	at org.apache.tika.parser.mp3.MpegStream.skipStream(MpegStream.java:283)
	at org.apache.tika.parser.mp3.MpegStream.skipFrame(MpegStream.java:160)
	at org.apache.tika.parser.mp3.Mp3Parser.getAllTagHandlers(Mp3Parser.java:193)
	at org.apache.tika.parser.mp3.Mp3Parser.parse(Mp3Parser.java:71)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.AutoDetectParser.parse(AutoDetectParser.java:120)
	at org.apache.tika.Tika.parseToString(Tika.java:380)
	...

"XWiki Solr index thread" daemon prio=10 tid=0x0000000003b72800 nid=0x64b5 runnable [0x00007f46f4618000]
   java.lang.Thread.State: RUNNABLE
	at org.apache.tika.io.TailStream.skip(TailStream.java:133)
	at org.apache.tika.parser.mp3.MpegStream.skipStream(MpegStream.java:283)
	at org.apache.tika.parser.mp3.MpegStream.skipFrame(MpegStream.java:160)
	at org.apache.tika.parser.mp3.Mp3Parser.getAllTagHandlers(Mp3Parser.java:193)
	at org.apache.tika.parser.mp3.Mp3Parser.parse(Mp3Parser.java:71)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.AutoDetectParser.parse(AutoDetectParser.java:120)
	at org.apache.tika.Tika.parseToString(Tika.java:380)
	...

"XWiki Solr index thread" daemon prio=10 tid=0x0000000003b72800 nid=0x64b5 runnable [0x00007f46f4617000]
   java.lang.Thread.State: RUNNABLE
	at java.io.BufferedInputStream.read1(Unknown Source)
	at java.io.BufferedInputStream.read(Unknown Source)
	- locked <0x00000000cb1be170> (a java.io.BufferedInputStream)
	at org.apache.tika.io.ProxyInputStream.read(ProxyInputStream.java:99)
	at java.io.FilterInputStream.read(Unknown Source)
	at org.apache.tika.io.TailStream.read(TailStream.java:117)
	at org.apache.tika.io.TailStream.skip(TailStream.java:140)
	at org.apache.tika.parser.mp3.MpegStream.skipStream(MpegStream.java:283)
	at org.apache.tika.parser.mp3.MpegStream.skipFrame(MpegStream.java:160)
	at org.apache.tika.parser.mp3.Mp3Parser.getAllTagHandlers(Mp3Parser.java:193)
	at org.apache.tika.parser.mp3.Mp3Parser.parse(Mp3Parser.java:71)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.CompositeParser.parse(CompositeParser.java:242)
	at org.apache.tika.parser.AutoDetectParser.parse(AutoDetectParser.java:120)
	at org.apache.tika.Tika.parseToString(Tika.java:380)
	...

This makes our Solr indexer very fragile as it prevents it from indexing other files thus leading to incomplete search results.

Issue Links

• is related to: TIKA-788 DWG parser infinite loop on possibly corrupt file
• links to: XWIKI-9528: A corrupt mp3 file can block the Solr indexing thread