Hi Jarek, Thank you for your reply. But sorry, I didn't understood all things :)
First you said that you have used the standalone jar file. After that you said that you have used the war. I agree with you, I prefer the third model, as you explained. My question, if we want to implement this model we must use the jar file? or we can use the provided war? In my opinion, the war is more secure because you benefit of the tomcat connection pooling. I don't know how we can manage that with a standalone jar. For me the the jar is only for basic use and testing, explained in [1]. But in [2], they say the "repository server" model is more used for production purpose. In the same page we see in the schema "standalone server content repository". Do they refer to the jar file? I think I am confusing the standalone jar and the third deployment model (repository server). In your case, could you please explain me more what was your final architecture: - How many repository you use? is it shared? - How do you expose your repository? jar or war? and how many of them? - How type of repository consumers you have? In the meantime, I have tested the CLI tool of jackrabbit [3], it works well. I can read and write on the repository. [1] http://jackrabbit.apache.org/standalone-server.html [2] http://jackrabbit.apache.org/deployment-models.html [3] http://wiki.apache.org/jackrabbit/InteractiveCommandLine Thanks, Tarek On Sat, Oct 4, 2014 at 11:08 PM, Jaroslaw Marek <jaroslaw.ma...@gmail.com> wrote: > Hi Tarek, > I assume you've seen Jackrabbit documentation [1]. We use the last model - > Standalone Repository Server - for two reasons: > > 1. We can scale cluster of applications servers and cluster of > repository servers independently. > 2. More than one application can use the same repository. > > What we did is we took jackrabbit-webapp and we deployed it in a Tomcat > configured by us and "inject" our own repository.xml configuration [2] (it > is different than using jackrabbit-standalone.jar). > > Other thing is how you want to connect to your repository. In case of > standalone server there are two options: > > * RMI > * WebDAV > > Whichever protocol you choose, you can either use JCA from within your > application or create Repository object manually. > > Coming to your question about administrating the repository - as far as I > know there is no perfect tool for that. I tried Toromiro [3] and JCR > Explorer [4], both have pros and cons. I haven't tried CLI tool yet, but I > would imagine it would be possible to use that. I will try. > > Cheers, > Jarek > > [1] http://jackrabbit.apache.org/deployment-models.html > [2] http://wiki.apache.org/jackrabbit/Clustering > [3] http://www.subshell.com/en/toromiro/jcr-browser-editor/ > [4] http://www.jcr-explorer.org/ > > > On 03.10.2014 23:00, Tarek Jarraya wrote: > >> Hi All, >> >> We want to use jackrabbit 2.8 with the following needs: >> - We use only servlet container Tomcat 7. Maybe it is a constraint for JCA >> adapter deployment mode. >> - We want to be able to have access to administrate the repository : multi >> access channels >> >> The only mode that can match is the standalone. But based on the wiki >> description is more adapted for quick start not for real production >> purpose. >> I have deployed the war version on tomcat, but unfortunately I can not >> access to it with the CLI tool based on the standalone jar. >> >> Which deployment model do you advice me? >> >> Thank you, >> Tarek >> >> >
