Shahzad - I am curious what features of distributed search stops you to run SolrCloud. Using DS, you would be able to search across cores or collections. https://cwiki.apache.org/confluence/display/solr/Advanced+Distributed+Request+Options
Thanks, Susheel On Tue, Feb 9, 2016 at 12:10 AM, Shahzad Masud < shahzad.ma...@northbaysolutions.net> wrote: > Thank you Shawn for your response. I would be running some performance > tests lately on this structure (one JVM with multiple cores), and would > share feedback on this thread. > > >There IS a way to specify the solr home for a specific context, but keep > >in mind that I definitely DO NOT recommend doing this. There is > >resource and administrative overhead to running multiple copies of Solr > >in one JVM. Simply run one context and let it handle multiple shards, > >whether you choose SolrCloud or not. > Due to distributed search feature, I might not be able to run SolrCloud. I > would appreciate, if you please share that way of setting solr home for a > specific context in Jetty-Solr. Its good to seek more information for > comparison purposes. Do you think having multiple JVMs would increase or > decrease performance. My document base is around 20 million rows (in 24 > shards), with document size ranging from 100KB - 400 MB. > > SM > > On Mon, Feb 8, 2016 at 8:09 PM, Shawn Heisey <apa...@elyograg.org> wrote: > > > On 2/8/2016 1:14 AM, Shahzad Masud wrote: > > > Thank you Shawn for your reply. Here is my structure of cores and > shards > > > > > > Shard 1 = localhost:8983/solr_2014 [3 Core - Employee, Service > Tickets, > > > Departments] > > > Shard 2 = localhost:8983/solr_2015 [3 Core - Employee, Service > Tickets, > > > Departments] > > > Shard 3 = localhost:8983/solr_2016 [3 Core - Employee, Service > Tickets, > > > Departments] > > > > > > While searching, I use distributed search feature to search data from > all > > > three shards in respective cores e.g. If I want to search from Employee > > > data for all three years, I search from Employee core of three > contexts. > > > This is legacy design, do you think this is okay, or this require > > immediate > > > restructure / design? I am going to try this, > > > > > > Context = localhost:8982/solr (9 cores - Employee-2014, Employee-2015, > > > Employee-2016, ServiceTickets-2014, ServiceTickets-2015, > > > ServiceTickets-2016, Department-2014, Department-2015, Department-2016] > > > distributed search would be from all three cores of same data category > > > (i.e. For Employee search, it would be from Employee-2014, > Employee-2015, > > > Employee-2016). > > > > With SolrCloud, you can have multiple collections for each of these > > types and alias them together. Or you can simply have one collection > > for employee, one for servicetickets, and one for department, with > > SolrCloud automatically handling splitting those documents into the > > number of shardsthat you specify when you create the collection. You > > can also do manual sharding and split each collection on a time basis > > like you have been doing, but then you lose some of the automation that > > SolrCloud provides, so I do not recommend handling it that way. > > > > > Regarding one Solr context per jetty; I cannot run two solr contexts > > > pointing to different data in Jetty, as while starting jetty I have to > > > provide -Dsolr.solr.home variable - which ends up pointing to one data > > > folder (2014 data) only. > > > > You do not need multiple contexts to have multiple indexes. > > > > My dev Solr server has exactly one Solr JVM, with exactly one context -- > > /solr. That instance of Solr has 45 indexes (cores) on it. These 45 > > cores are various shards for three larger indexes. I am not running > > SolrCloud, but I certainly could. > > > > You can see 25 of the 45 cores in my Solr instance in this screenshot of > > the admin UI for this server: > > > > https://www.dropbox.com/s/v87mxvkdejvd92h/solr-with-45-cores.png?dl=0 > > > > There IS a way to specify the solr home for a specific context, but keep > > in mind that I definitely DO NOT recommend doing this. There is > > resource and administrative overhead to running multiple copies of Solr > > in one JVM. Simply run one context and let it handle multiple shards, > > whether you choose SolrCloud or not. > > > > Thanks, > > Shawn > > > > >
