I would _really_ advise you avoid trying this without an absolutely compelling, _demonstrated_ need.
Unless you have a _lot_ of cores, the load time for even 100 or so isn't horrible, so if you're noticing a significant lag time, it's probably worth looking at why that is before jumping into using unsupported options. If you insist on doing this, I'll be anxiously awaiting reports of whatever problems you have. There are tests that hammer this functionality, see TestLazyCores and the speed that cores hums right along. So the first thing I'd look at is if your configuration is such that you have large transaction logs that are getting replayed at startup, see: http://wiki.apache.org/solr/SolrPerformanceProblems#Slow_startup After that I'd try to pin down why you are seeing long load times if that's not relevant. And you haven't told us anything about your setup, this smells like an XY problem. Best Erick On Fri, Aug 30, 2013 at 10:13 AM, Srivatsan <ranjith.venkate...@gmail.com>wrote: > I started looking in reducing the time taken to load cores during cluster > restart. For initializing the core, building config file takes considerable > amount of time. In our case, schema remains same for all collections(cores) > . Hence it can be kept static, and avoid loading time. > > In the mean while i have gone through schemaless feature of solr-4.5. In > such a case, do we need to define schema.xml file or it ll automatically > assigns field types dynamically? If there is no need to define schema.xml > file, then automatically core initialization takes lesser time rite? > > > Thanks in advance > > Srivatsan > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Collection-loadOnStartup-tp4082531p4087510.html > Sent from the Solr - User mailing list archive at Nabble.com. >
