Thank you Erik and Shawn for your support. I'm using Solr's Schema API and Config API to manage and administer a Solr deployment based on customer specific setting that my application will need to do to a Solr deployment. A client application will be using my APIs and as part of data validation, I'm trying to figure out what to allow and what not too as invalid attributes data that I cannot send to Solr.
For example, I wasn't sure that a request-handler name can have spaces or can be all numeric, etc. What about fields and field types, is there a restriction for the field names? I know my question is broad, but if there is a starting point, I can use that to help me write application so that it is defensive against clients who will use it to manage Solr. If they use invalid data, I don't want to send it to Solr and cause Solr to break. Steve On Fri, Jun 12, 2015 at 4:41 PM, Erik Hatcher <erik.hatc...@gmail.com> wrote: > Do note that AdminHandler*s* (plural) is "* A special Handler that > registers all standard admin handlers”, so if you’re trying to do something > tricky with admin handlers, Note that AdminHandlers is also deprecated and > these admin handlers are implicitly registered with ImplicitPlugins these > days. > > What kind of handler are you adding? Or are you trying to change the > prefix of all the admin handlers instead of /admin? > > — > Erik Hatcher, Senior Solutions Architect > http://www.lucidworks.com > > > > > > On Jun 12, 2015, at 3:56 PM, Shawn Heisey <apa...@elyograg.org> wrote: > > > > On 6/12/2015 1:02 PM, Steven White wrote: > >> You are right. If I use "solr.SearchHandler" it works, but if I > >> use "solr.admin.AdminHandlers" like so: > >> > >> <requestHandler name="987" class="solr.admin.AdminHandlers"> > >> </requestHandler> > >> > >> Solr reports this error: > >> > >> HTTP ERROR 500 > >> > >> Problem accessing /solr/db/config/requestHandler. Reason: > >> > >> {msg=SolrCore 'db' is not available due to init failure: The > >> AdminHandler needs to be registered to a path. Typically this is > > > > With an admin handler, it won't be possible to access that handler if > > it's not a path that starts with a slash, so it can be incorporated into > > the request URL. Solr is making sure the config will work, and throwing > > an error when it won't work. > > > > With search handlers, if you set up the config right, you *CAN* access > > that handler through a request parameter on the /select handler, it does > > not need to be part of the URL path. The default config found in > > examples for 4.x and later will prevent that from working, but you can > > change the config to allow it ... so search handlers must work even if > > the name is not a path. The config validation is not as strict as it is > > for an admin handler. > > > > Thanks, > > Shawn > > > >
