I found a blog a few days ago (don't remember the source now) stating that if you name the index folder as "index" and not "index.xxxxxxxxxxxxx", then the slave would stop pulling the full catalog every time.
That was what I did and that seemed to be a good workaround for me. I stopped solr, renamed index.xxxxxxxx folder to just index and modified index.properties file to use the index folder. #index.properties #Thu Feb 21 02:29:09 CST 2013 #index=index.20130221022400014 index=index master->slave replication seemed to behave as expected after this change for me. Hung -----Original Message----- From: Mark Miller [mailto:markrmil...@gmail.com] Sent: Wednesday, February 27, 2013 11:09 AM To: solr-user@lucene.apache.org Subject: Re: Solr 4.1 Master -> Slave Replication Issues and Solr 4.2 On Feb 27, 2013, at 10:51 AM, Michael Della Bitta <michael.della.bi...@appinions.com> wrote: > Are there any temporary fixes to this problem aside from waiting for 4.2? Not really. I guess my advice would be this: If you use SolrCloud or single node, don't worry, this doesn't affect you. If you use 4.1 master->slave replication and you have not noticed this, I'd just wait for 4.2. There is not a lot you can do to work around it. You might ease up on how often you poll for changes if there is a problem. If you need something that doesn't replicate so often right now, 4.0 does not have the same issue. There might be the rare occasion it replicates when it does not need to, but it's not consistent as in 4.1. How did this happen? There was not a test to make sure it didn't happen and from a functionality point, it doesn't mess anything up for tests to detect. Because of that, other changes introduced this problem and it was not detected nor noticed by users until recently. I've added testing around it with the fixes for the future. - Mark
