On 25/08/11 12:43 AM, Christopher Schultz wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Aristedes,
On 8/24/2011 10:36 AM, Aristedes Maniatis wrote:
On 25/08/11 12:15 AM, Christopher Schultz wrote:
So, if you could do this with code it would be:
getServletContext().getRealPath("/WEB-INF/classes/solr")
Right?
Yes.
Okay... glad I understand the goal :)
There isn't any system property for this kind of thing (since
system properties are JVM-wide, and not specific to any
particular webapp) so it would have to be, as you say, some kind
of "magic" variable that you could use.
Are all properties JVM-wide?
All system properties (retrieved using System.getProperty()) are
JVM-global.
If I define a property in:
webapps/someapp##1234/META-INF/context.xml
Those aren't system properties -- they /will/ be local to the webapp.
Right, then theoretically at least, there could be a ${APP_ROOT} which could be
available in there so that paths within the application could be addressed. I
guess this use case is rare since code within the application is usually able
to find resources on the classpath within the application.
I don't believe such a variable exists. Since Solr is servlet
context-aware application (i.e. it *is* webapp), is it possible
to give it a path relative to the webapp itself? You might want
to ask the Solr folks about this.
Sure. I can try there next, but any such solution is going to
involve changing Solr code.
Maybe not. Most webapp configurations assume that paths are relative
to the webapp itself.
Our testing so far seems to suggest that it is relative to the tomcat install.
But I'll follow this up with the Solr people. Thanks for the pointer.
Another option, which I think is a better option, would be to
locate solr/home outside of your deployment directory.
That is an option, but a very awkward one since then I have an
application separated from the configuration which is an integral
part of its functionality.
What goes in solr/home, anyway? I figured that the indexes go in
there, not just a configuration file or two.
It contains files such as the schema definition (which is similar to a database
schema definition, only for Lucene), instructions on how to parse data sources,
weightings for query facets, etc. The actual index can be in a different path.
This isn't some JDBC connection which differs in every deployment,
but the configuration of exactly how the application works. So
being able to keep it in our SCM with the code and deploy together
is essential.
Fair enough. I think the Solr list is the place to go. Definitely let
us know what the answer is.
Thanks again
Ari
--
-------------------------->
Aristedes Maniatis
GPG fingerprint CBFB 84B4 738D 4E87 5E5C 5EFA EF6A 7D2E 3E49 102A
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
