jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexander Klimetschek <aklim...@adobe.com>
Subject Re: Question regarding Jackrabbit clustering
Date Sun, 07 Aug 2011 16:15:18 GMT
(answering to the Jackrabbit users mailing list, which you might want to join)

A persisted Jackrabbit repository on the file system can only be used by a single Jackrabbit
instance. If you have 2 different JVM processes, they need to either have their own cluster
node on the filesystem (although that usually only makes sense if they are on different machines)
or they communicate over some remoting, be it HTTP (use Apache Sling for that) or some of
the Jackrabbit remoting using WebDAV or RMI.


On 22.07.11 08:28, "Nune Isabekyan" <nisabek@gmail.com<mailto:nisabek@gmail.com>>


I'm sorry for bothering you, but I have a question regarding Jackrabbit clustering and I have
found you email in the web and your answers regarding some problems of Jackrabbit.

I would really appreciate your help.

So the question is: As far as I understood the clustering works this way: each cluster has
it's own repository and workspace, but they are configured to use the same persistence place
(like DB) and are being sync. But the problem is that Jackrabbit locks repository (creating
.lock file and doesn't let another process (say another tomact) connect to it). SO clustering
does not basically solve the problem of concurrent read from different processes? I mean we
have to use separate node for each process?

And if yes are there ANY options for accessing jackrabbit from several processes? except RMI
of course.

Thanks a lot in advance.

Software Engineer at be2 Germany GmBH

Alexander Klimetschek
Developer // Adobe (Day) // Berlin - Basel

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message