Walkthrough

The following steps simulate a typical user session. An end result (i.e a layout of file and directories) is then shown.

...

Add Content

curl -vX -XPUTPUT --uploaddata-filebinary fcrepo4"@fcrepo4_greetings.txt" "http://localhost:8080/rest/greetings_en/ds0/fcr:content"

Fedora will create a directory "fcrepo4-data" in the current working directory. The default directories found in "fcrepo4-data" will be the following:

...

The serialized Fedora nodes can be found in the "fcrepo.ispn.repo.CacheDirPath/FedoraRepository" directory. Since, the user specified X configuration (as opposed to Y, e.g.), the The files in that directory would look something like this:

...

Some key things to note about the files and their contents:

• The file names in the case of FileCacheStore are just signed 32-bit integers, so the usual and maximum length of file names is 10 or 11 characters. The file names (i.e. integers) are generated by calculating the Java

  hashcode

  hashCode of the node's UUID string (prefixed with root node UUID and

  shifting

  masking it by a 22 fixed bit mask). For example, the following Java snippet code calculates the file name for root node (UUID "87a0a8c75085d64/"):

  Code Block
  language java
  fileName = "87a0a8c7505d64/".hashCode() & 0xfffffc00; //equals -891938816

  Therefore, file -891938816

   gets generated, containing

   would contain data about the root node

  children

  . 
  
  

• Each of the files contain serialized ModeShape nodes. Although the generated files are binary, the data can be read using a tool that understands both BSON and JBoss serialization, or by using ModeShape API itself (It does not seem possible using a tool (a simple utility is under development) can help in reading the contents of the file (It might not be easy to read these files using existing bson tools, like mongoDB bsondump). Using a tool such as X can help in reading the contents of the file alone). For example, for our root node file (-891938816) , a tool could show might display the root node data in JSON as following:

  Code Block
  language js
  {  "properties" :  {  "http://www.jcp.org/jcr/1.0" :  { "primaryType" :  { "$name" : "mode:root" } , "uuid" : "87a0a8c7505d64/"  }  } , "children" :  [ { "key" : "87a0a8c317f1e7jcr:system" , "name" : "jcr:system" } ,  { "key" : "87a0a8c7505d646988a17a7c366d9-ad2f25bc-48f64e2d-aef59e53-e7d411e184d94428fe7b8152" , "name" : "chandnigreetings_en" } ] , "childrenInfo" : { "count" : 2 }   }

  As an aside, X would a tool could be written to show the full file-to-node mapping ascontent mapping; the utility under development aims to display this information like the following snippet: 

  Code Block
  -891938816 UUID: 87a0a8c7505d64 mode:root 1036310528-1857312768 UUID: 7c366d9-25bc-4e2d-9e53-4428fe7b8152 greetings_en -112221211 UUID: 328781824 UUID 7e84184d-3a8e-4f57-8e49-a550f1c19b3a greetings_en/ds0 39043072 UUID fc859bab-6f7b-46de-9b4c-f40dfe28643c hello, world!

• The serialization is done by a the JBoss serialization library, not JDK's native object serialization machinery. For this reason the generated serialized files look different from an ordinary JDK serialized file. JBoss Marshalling can be configured to use custom serialization classes that read and write content in the format of the repository's choosing. 
  
  

• The data is encoded in Binary JSON (BSON).

  (Please refer to section on node contents for details.)

  If the file containing the root node (referencing the node 'greetings_en') is opened up in a hex editor, you would see /u0002 preceding strings (such as "name","key"); /u0004 preceding an array (representing sub-nodes); /u0003 representing the UUID -- in accordance with the BSON spec.

  Code Block

  \u0000\u0000\u0000\u0004children\u0000ü\u0000\u0000\u0000\u00030\u0000<\u0000\u0000\u0000 \u0002key\u0000\u0019\u0000\u0000\u000087a0a8c317f1e7jcr:system\u0000\u0002name\ u0000\u000B\u0000\u0000\u0000jcr:system\u0000\u0000\u00031\u0000X \u0000\u0000\u0000\u0002key\u00003\u0000\u0000\u000087a0a8c7505d6417c366d9-25bc-4e2d-9e53-4428fe7b8152\u0000\u0002name\u0000\u0000\u0000\u0000greetings_en\u0000\u0000\u0000\u0003childrenI

Inspecting Individual Binary Data Files

Key elements are annotatedSome of these elements are:

1. 87a0a8c7505d64 refers to root node UUID

2. The grayed out text refers to Modeshape classes responsible for data representation and serialization. The node serialization is done by JBoss Marshalling , which can be configured to use custom serialization classes that read and write content in the format of their choosing (in this case BSON). The serialization format contains the name of datastructures and ModeShape i.e. custom marshallers (classes in org.infinispan.schematic.internal.*) 

3. Node name.

4. The file contains binary data for specifying different internal attributes, so when opened up in

   a

   an ordinary text editor, the

   file

   text editor might show some numbers as garbage characters

   or numbers in strings. /children0<key87a0a8c317f1e7jcr:systemnamejcr:system1Skey387a0a8c7505d64 . Similarly, if object 'chandni' had siblings, the array would appear in the binary as . . .:Node children UUID and name (6988a17a-ad2f-48f6-aef5-e7d411e184d9namechandni)

   . These numbers have meaning for node's attributes, though. For example, the highlighted entry #4 specifies an array of the children of a node. The index number of the child precedes its key and name. So, if object "greetings_en" gets another sibling "greetings_fr", the latter's entry would be preceded by 2 (its index). The full children node array would now appear something like: 

   /children0<key87a0a8c317f1e7jcr:systemnamejcr:system 1Xkey387a0a8c7505d6417c366d9-25bc-4e2d-9e53-4428fe7b8152namegreetings_en 2Xkey387a0a8c7505d64aa21b2ee-d22a-4b0a-8b43-a8e8b58c5ec6namegreetings_fr

5. Children count in binary.

...

Similarly, for the datastream content (omitting ModeShape API artifacts):

Some of the The elements of interests are:

1. Root UUID
2. Omitted Serialization artifacts (see note on parent node for details).
3. UUID
4. Content type of document (in this case it's the default content type for documents). See 8
5. .
7. Parent UUID (datastream)
8. Actual content ("hello world" text in this case)Actual text content.
9. Last modified date.
10. Last modified admin. 
11. Content type of datastream.
12. SHA-1 generated by Fedora.
    

The corresponding readable representation is:

...

1. .

...

Inspecting Indexing Folder

ModeShape currently makes use of Hibernate Search to manage Lucene indexes. The indexes can be viewed by using Luke, e.g.
These files can be found in the "fcrepo.modeshape.index.location" directory. 

Inspecting ObjectStore Folders

Directories "com.arjuna.ats.arjuna.objectstore.objectStoreDir" and "com.arjuna.ats.arjuna.common.ObjectStoreEnvironmentBean.default.objectStoreDir" are JBoss JTA transaction engine artifacts. The default Fedora Infinispan configuration attempts to find a JBossJTA transaction manager implementation via   "org.infinispan.transaction.lookup.GenericTransactionManagerLookup". This configuration uses Arjuna ShadowFileStore as a backend, resulting in several directories within fcrepo4-data such as "object-store" and "object-store-default":

   |-object-store
   |---ShadowNoFileLockStore
   |-----defaultStore
   |-------Recovery
   |---------TransactionStatusManager
   |-object-store-default
   |---ShadowNoFileLockStore
   |-----defaultStore

A detailed description of the artifacts maintained by the JBossJTA implementation is most likely beyond the scope of this document (at least for now).

Infinispan Configuration Options

Depending on the configured Infinispan backend, the directory layout and contents of the binary files would be different. The following sections covers other cache store options.

...

LevelDB Backend

Currently, the default configuration outputs Fedora data to LevelDB (a fast filesystem based key-value store). When Fedora 4 is started, ModeShape (actually Infinispan and LevelDB in the background) will create several directories on the filesystem. Currently, the directories created are:

...

1. File .log holds entries for recent transactions. The relevant API for representing these entries is modeshape-schematics (see, e.g.,  org.infinispan.schematic.SchematicEntry)
2. File .sst stores these entries when the .log file reaches a size threshold. A new log file is generated.
3. File MANIFEST.x records info about .sst files (among other things). 
4. File CURRENT specifies the current MANIFEST file.

Most of these files are binary and can be read by a LevelDB Java library.

Inspecting Individual Binary Data Files

As is the case with the FileCacheStore, the .log files are binary. The default initial .log file is 000003.log, and it contains serialized entries for all the nodes from recent transactions. If the repository has only a few hundred added nodes, this file will contain all the nodes (i.e. the additions and default properties like mix:etag).