version storage question

[Giota Karadimitriou <Gi...@eurodyn.com>]

I would like to ask a general question regarding how version storage is
organized in jackrabbit.
 
>From what I see for each version node in jackrabbit moving up from
parent to parent there is:
 
a)       A parent version history node
b)       parent of a is a version storage node
c)       parent of b is another version storage node
d)       parent of c is another version storage node
e)       parent of d is root
 
so it's like: 
root-->d(version storage)-->c(version storage)-->b(version
storage)-->a(version history)-->version_node
 
why are there 3 version storages? 
 
'b' has 'a' as its only child but c does not have b as its only
child.what are these other children?
 
In general what do the second and third version storage nodes (c and d
in my example) contain?
 
Regards
giota
 
 

Re: version storage question

[Tobias Bocanegra <to...@day.com>]

when you look at the code of createVersionHistory() in AbstractVersionManager:

[...]
            // create deep path
            String uuid = node.getNodeId().getUUID().toString();
            NodeStateEx root = historyRoot;
            for (int i = 0; i < 3; i++) {
                QName name = new QName(QName.NS_DEFAULT_URI,
uuid.substring(i * 2, i * 2 + 2));
                if (!root.hasNode(name)) {
                    root.addNode(name, QName.REP_VERSIONSTORAGE, null, false);
                    root.store();
                }
                root = root.getNode(name, 1);
            }
            QName historyNodeName = new QName(QName.NS_DEFAULT_URI, uuid);
[...]


you see that the path to the version history is built using the uuid
of the version history node. the idea is to create a good distribution
of the histories in the version storage. this is mainly because
jackrabbit can't handle large number of childnodes very well.

regards, toby

On 6/28/06, Giota Karadimitriou <Gi...@eurodyn.com> wrote:
> I would like to ask a general question regarding how version storage is
> organized in jackrabbit.
>
> From what I see for each version node in jackrabbit moving up from
> parent to parent there is:
>
> a)       A parent version history node
> b)       parent of a is a version storage node
> c)       parent of b is another version storage node
> d)       parent of c is another version storage node
> e)       parent of d is root
>
> so it's like:
> root-->d(version storage)-->c(version storage)-->b(version
> storage)-->a(version history)-->version_node
>
> why are there 3 version storages?
>
> 'b' has 'a' as its only child but c does not have b as its only
> child.what are these other children?
>
> In general what do the second and third version storage nodes (c and d
> in my example) contain?
>
> Regards
> giota
>
>
>
>


-- 
-----------------------------------------< tobias.bocanegra@day.com >---
Tobias Bocanegra, Day Management AG, Barfuesserplatz 6, CH - 4001 Basel
T +41 61 226 98 98, F +41 61 226 98 97
-----------------------------------------------< http://www.day.com >---