Sverker Abrahamsson created CAMEL-9735:
------------------------------------------
Summary: camel-tarfile throws OutOfMemoryError when splitting
large files
Key: CAMEL-9735
URL: https://issues.apache.org/jira/browse/CAMEL-9735
Project: Camel
Issue Type: Bug
Components: camel-tarfile
Affects Versions: 2.16.2
Reporter: Sverker Abrahamsson
After the change in https://issues.apache.org/jira/browse/CAMEL-9375 I started
to get failures with the following stacktrace:
{noformat}
Caused by: java.lang.OutOfMemoryError: Java heap space
at java.util.Arrays.copyOf(Arrays.java:3236)
at java.io.ByteArrayOutputStream.grow(ByteArrayOutputStream.java:118)
at
java.io.ByteArrayOutputStream.ensureCapacity(ByteArrayOutputStream.java:93)
at java.io.ByteArrayOutputStream.write(ByteArrayOutputStream.java:153)
at
org.apache.camel.converter.stream.CachedOutputStream.write(CachedOutputStream.java:104)
at org.apache.camel.util.IOHelper.copy(IOHelper.java:199)
at org.apache.camel.util.IOHelper.copy(IOHelper.java:169)
at org.apache.camel.util.IOHelper.copyAndCloseInput(IOHelper.java:218)
at org.apache.camel.util.IOHelper.copyAndCloseInput(IOHelper.java:214)
at
org.apache.camel.converter.stream.StreamCacheConverter.convertToStreamCache(StreamCacheConverter.java:82)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:497)
at
org.apache.camel.util.ObjectHelper.invokeMethod(ObjectHelper.java:1243)
at
org.apache.camel.impl.converter.StaticMethodTypeConverter.convertTo(StaticMethodTypeConverter.java:59)
at
org.apache.camel.impl.converter.BaseTypeConverterRegistry.doConvertTo(BaseTypeConverterRegistry.java:293)
at
org.apache.camel.impl.converter.BaseTypeConverterRegistry.mandatoryConvertTo(BaseTypeConverterRegistry.java:168)
at
org.apache.camel.dataformat.tarfile.TarIterator.createNextMessage(TarIterator.java:121)
at
org.apache.camel.dataformat.tarfile.TarIterator.tryAdvanceToNext(TarIterator.java:100)
at
org.apache.camel.dataformat.tarfile.TarIterator.hasNext(TarIterator.java:78)
at
org.apache.camel.processor.Splitter$SplitterIterable$1.hasNext(Splitter.java:163)
at
org.apache.camel.processor.Splitter.updateNewExchange(Splitter.java:256)
{noformat}
I.e. using StreamCache makes the whole entry go into memory and since they in
my usecase are several gigabytes it doesn't work very well. I realize that this
might be caused since I don't set any StreamCachingStrategy on my context but
reading the corresponding doc part I get the impression that the default
settings should be that it will spool to disk after reading 128 kbyte.
However, I really don't want it to do that neither as that would 1) be a
significant overhead to write gigabytes of data to a temp file and 2) I might
not even have that much space where my temp directory is located. Instead the
functionality that I want is that the tarfile entries are streamed and written
directly to the destination files without any unnecessary caching in between.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
