Independent/Selfcontained Solr Unit testing with JUnit
Hi, Is there any way to run self-contained JUnit tests for say a Solr dependent class where it doesn't depend on Solr being up and running at localhost:8983 ? I have a collection etc. setup on the Solr server. Is it possible to mockit with an EmbeddedSolr easily with a @Before or @BeforeClass annotation in JUnit4 ? Any pointers to examples would be awesome(I am also trying to look in the source). TIA, Vijay
Swap space,JVM-Memory,Physical memory on Solr Admin UI explanation
Hi, I am using solr 4.6.1 on a Windows 7 server right now with 32 GB RAM.I have a SolrCloud with 3 shards, 2 replicas and an embedded Zookeeper on the 1 box.I have allocated -Xmx5GB RAM to each Solr instance when starting up with -XX:MaxNewSize:1636m I see the Swap space(32.5G/64GB),JVM-Memory(521.1MB/4.73GB),Physical memory(11.07 GB/32GB) on Solr Admin UI. That usage is confusing me. The Swap space was going up when indexing 15 million documents but not the JVM-memory(which went up a max of 1.1G or so). So, does that mean I don't need to allocate that much RAM for each Solr instance ? Could someone explain the 3 terms clearly in terms of their use in Indexing and Querying: Swap space, JVM-memory and Physical memory ? TIA, Vijay
How do I use Solr for Search only instead of SQL Server while using RDBMS for write?
Hi, We use SQL Server currently.We are thinking about using technologies like Solr to speed up our search into the RDBMS-hopefully by flattening the RDBMS data into a denormalized data structure and indexed for the various queries performed by our users.This causes a mismatch between read-only search and subsequent write operation the user might perform on the returned data-an architectural smell.Right now, the search screen has a whole bunch of form fields to search to return the data from the RDBMS in a data grid below the search form.Any advise on how to handle this use case with Solr + Storm as an example (near real time of 30 mins)? Heard there is a plugin for Solr use with RDBMS's(DIH) ? Are they useful for this use case? My earlier work place had lots of issues with SQL server with regards to space. Even if we return from the denormalized dataSource through Solr, we need to provide a way for users to do Update data in the RDBMS using fields or a REST service ? The RDBMS data has to be denormalized to avoid making 14+ queries to SQL server.Pushing this data to a flat structure/dataSource for Solr to then use was what I was thinking about-not exactly real time and that is fine for our use case. Are there any AWS instances that can get me started on Solr + Storm ? TIA, Vijay
