Re: Replicator and Writer deletion policy

2017-06-01 Thread Riccardo Tasso
Thank you Michael, I'm waiting for the Shai answer. Unfortunately I can't move from lucene4 but I will take a look at the new NRT replication. 2017-06-01 12:34 GMT+02:00 Michael McCandless : > Sorry, I don't know the answer to your questions; maybe Shai Erera does? > > But, FYI: Lucene added nea

Re: searchAfter is missing results when custom noncontinuous slices are used

2017-06-01 Thread Christoph Kaser
Hello Mike, thank you for the explanation! I created a jira issue: LUCENE-7861 Best regards, Christoph Am 25.05.2017 um 16:11 schrieb Michael McCandless: Yes, there is a (hidden) assumption in TopDocs.merge that the hits it's merging are logically non-overlapping, sequential slices of the inde

Re: Replicator and Writer deletion policy

2017-06-01 Thread Michael McCandless
Sorry, I don't know the answer to your questions; maybe Shai Erera does? But, FYI: Lucene added near-real-time replication recently ( https://issues.apache.org/jira/browse/LUCENE-5438), for much faster replication (no commit needed). It also doesn't require you to use SnapshotDeletionPolicy. Ins

Replicator and Writer deletion policy

2017-06-01 Thread Riccardo Tasso
I am studying a solution which uses lucene replicator components (I'm on lucene 4.10.4) and I've found some documentation on the javadoc and in some blog posts. When my IndexWriter commits it produces a Revision which is sent to the ReplicationClient. To be sure of keeping data untill the replica