First, I'd be sure that it's worth the effort. Are you doing a massive number of updates? HTTP isn't all that time-consuming, so how many documents are you talking about here?
Doing this inside Solr will stress Solr as well, even in the atomic update world. Your poor Solr disk has to go out and collect the document first, then go through all the work of re-indexing it. It's quite possible that your overall throughput would be faster by creating a series of parallel clients and re-indexing from scratch. Bottom line. Be sure you need to do this before wasting the time. If your end-to-end indexing is fast enough, don't bother <G>.. Best Erick On Mon, Feb 4, 2013 at 9:58 AM, vittar <vital...@interlink-ua.com> wrote: > Hi! > I need to do some periodical updates in my Solr app(v.4.1). > There is some logic for this updates operation, and all the data I need for > this updates I already have in the index. Basically I have some small > collection which is synchronized with external out-side storage, and few > very big collections which I should update after small one is updated. > So I want to write some Solr plugin which will do my update logic inside > Solr stack, without losing time for http requests/responses. > Are there any suggestions how I can do this? Maybe I should write some > special DataImportHandler etc? Could somebody suggest something? > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Index-updates-inside-of-Solr-box-tp4038348.html > Sent from the Solr - User mailing list archive at Nabble.com. >
