jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marek Slama" <msl...@email.cz>
Subject Re: Versioning question
Date Fri, 19 Apr 2013 12:44:15 GMT
Ok. Thanks for answer. I checked for DB Data Store (Postgres) and it is the 
same ie. binary data are not duplicated.
(Actual db size can change but when I dump/restore it grows slowly ie. there
is small overhead for version history nodes only.)

I did simple test where I created many versions of the same file node when I
saved the same binary data/file to it (600KB and 2500 version
nodes).

Marek

-- 
Marek Slama
mslama@email.cz



---------- Původní zpráva ----------
Od: Daniel Hobi <d.hobi@gmx.ch>
Datum: 16. 4. 2013
Předmět: Re: Versioning question

"AFAIK, when using filedatastore, jackrabbit calculates a sha1 ckecksum over
the content of a file. The checksum will then be split up and used as 
folder/filename. 

If you have 2 identical files, the checksum will be exactly the same 
resulting in one file on the disk.

Daniel

Von meinem iPhone gesendet

Am 16.04.2013 um 15:34 schrieb "Marek Slama" <mslama@email.cz>:

> Hello,
> 
> I want to ask if Jackrabbit makes any space optimization when new version 
of
> binary content is created. For example:
> I have file node F: I create version V1 and then version V2 but I actually

> do not check in my code if anything has changed.
> so V1 1.0 and V2 1.1 content are the same. Is binary content stored twice 
or
> only once?
> 
> ie. my versionable node is:
> Node fileNode = documentNode.addNode(docxName, NodeType.NT_
> FILE);
> fileNode.addMixin(NodeType.MIX_VERSIONABLE);
> fileNode.addMixin(NodeType.MIX_CREATED);
> 
> Thanks
> 
> Marek
> 
> ="
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message