multiple instances of jackrabbit-standalone cause "file backing binary value not found" from org.apache.jackrabbit.util.TransientFileFactory

running 2 or more instances of jackrabbit-standalone causes file deletions in the temporary folder used by another standalone instance. (when garbage collected)

To reproduce, run 2 or more instances, create files in each, and then stop one of them and attempt read cached files by the other. The one that stopped will garbage collect files used by the other. This may be hard to reproduce, as it doesn't seem to be guaranteed to have a collision on file names. The problem "went away" when I forced each instance to use a different temporary folder. But this is not a permanent solution.
Ex:
java -Dhostname=standalonejcr -Djava.io.tmpdir=/tmp1 -Xmn100M -Xms500M -Xmx500M -jar jackrabbit-standalone-2.0.0.jar -p 8000 -r jcr-repository
java -Dhostname=standalonejcr -Djava.io.tmpdir=/tmp2 -Xmn100M -Xms500M -Xmx500M -jar jackrabbit-standalone-2.0.0.jar -p 8001 -r jcr-repository

Original Emails: (to jackrabbit dev mailing list)
>>>>>>>>>>>
Wed, Apr 7, 2010 at 11:21 AM
subject clustered environment, 2 different jvms, TransientFileFactory, storing file blobs in db

Hello,

I would normally file a bug on jira, but its very difficult to setup/reproduce, so I'm looking for insight first on how temp files/blobs are implemented in jackrabbit.

We currently run 2 different "standalone" instances of jackrabbit version 2.0.0, each in their own JVM and setup the same way in using <cluster>.

Our application connects to one of the standalone instances remotely(webdavex) for authoring content, and sends publish instructions (via JMS/activemq) to the other.

The problem though, is that BLOBInTempFile.getStream is occasionaly throwing : "file backing binary value not found", and one of the instances sometimes can't read the file.

I've searched and found this information:
http://mail-archives.apache.org/mod_mbox//jackrabbit-dev/200603.mbox/<90a8d1c00603150237t4c81df4fx178fcd726a93fe@mail.gmail.com>

So apparently, when files are read/written, you create a temporary cache, but TransientFileFactory runs as a singleton within a single JVM correct?

So can I assume that one of the "singletons", (there will be 2??) will delete files that were created by the other at some DIFFERENT random time when the garbage collector runs?

I've also attached your Repository.xml that we use for both (with different cluster ids of course)

Adrien

Thanks
Is there some way to avoid this??

I've attached our repository.xml for you to look at, both are setup the same way for e

Thanks.

>>>>>>>>>>>>
from Stefan Guggisberg
reply-to dev@jackrabbit.apache.org
to dev@jackrabbit.apache.org
date Thu, Apr 8, 2010 at 12:59 AM
subject Re: clustered environment, 2 different jvms, TransientFileFactory, storing file blobs in db
On Wed, Apr 7, 2010 at 8:21 PM, Adrien Lamoureux
wrote:
> Hello,
> I would normally file a bug on jira, but its very difficult to
> setup/reproduce, so I'm looking for insight first on how temp files/blobs
> are implemented in jackrabbit.
> We currently run 2 different "standalone" instances of jackrabbit version
> 2.0.0, each in their own JVM and setup the same way in using <cluster>.
> Our application connects to one of the standalone instances
> remotely(webdavex) for authoring content, and sends publish instructions
> (via JMS/activemq) to the other.
> The problem though, is that BLOBInTempFile.getStream is occasionaly throwing
> : "file backing binary value not found", and one of the instances sometimes
> can't read the file.
> I've searched and found this information:
> http://mail-archives.apache.org/mod_mbox//jackrabbit-dev/200603.mbox/<90a8d1c00603150237t4c81df4fx178fcd726a93fe@mail.gmail.com>
> So apparently, when files are read/written, you create a temporary cache,
> but TransientFileFactory runs as a singleton within a single JVM correct?

yes

> So can I assume that one of the "singletons", (there will be 2??) will
> delete files that were created by the other at some DIFFERENT random time
> when the garbage collector runs?

no, unless java.io.File#createTempFile invoked from 2 different jvm's
would create
colliding temp files. but that's impossible according to the javadoc [0]:

<quote>
[...] is guaranteed that:
1. The file denoted by the returned abstract pathname did not exist
before this method was invoked
[...]
</quote>

cheers
stefan

[0] http://java.sun.com/javase/7/docs/api/java/io/File.html#createTempFile(java.lang.String,
java.lang.String, java.io.File)

>>>>>>>>>>>>
from Thomas Müller
reply-to dev@jackrabbit.apache.org
to dev@jackrabbit.apache.org
date Thu, Apr 8, 2010 at 1:52 AM
subject Re: clustered environment, 2 different jvms, TransientFileFactory, storing file blobs in db

Hi,

Stefan is right, File.createTempFile() doesn't generate colliding
files. However, there is a potential problem with the
TransientFileFactory. Consider the following case:

• The file "bin-1.tmp" is created (BLOBInTempFile line 51).
• The TransientFileFactory adds a PhantomReference "A" in its queue.
• BLOBInTempFile.delete() or dispose() is called, the file "bin-1.tmp"
  is deleted.
• A new file is created, and also called "bin-1.tmp" is created
  (BLOBInTempFile line 51)
  (that's possible because File.createTempFile can re-use file names).
• The TransientFileFactory adds a second PhantomReference "B" in its
  queue, pointing
  to a different file with the same name.
• The first (only the first) BLOBInTempFile is no longer referenced.
• The TransientFileFactory.ReaperThread gets PhantomReference "A" and
  deletes this file. But the file is still used and referenced ("B").

I'm not sure if this is what is happening in your case, but it is a
potential problem.

Could you log a bug?

There are multiple ways to solve the problem. I think the best
solution is to not use File.createTempFile() and instead use our own
file name factory (with a random part, and an counter part).

Regards,
Thomas