Well, you have a couple of issues here: 1> the nodes shard ID will have to be guaranteed to be the same as it was when the shard was created. 2> you should be able to just copy the index directory to wherever your node is looking for it now. or 3> instanceDir should point to the directory that has subdirs of "conf" and "data". There is an additional parameter you can specify dataDir that should point to the actual index files (actually the directory that _contains_ the directory "index"
Best Erick On Tue, Nov 6, 2012 at 3:20 PM, Rasmussen, Chris <crasmus...@mitre.org>wrote: > I have a five node SolrCloud implementation running as a test with no > replication using a three node zookeeper ensemble. Admittedly, I'm new to > Solr and just grinding it out. Accidently re-initialized zookeeper with > the wrong conf dir and I'm trying to recover. I re-ran the initialization > with the correct conf dir, but now the indexes are reporting 0 documents. > Logs also report that a new index was created in the dataDir called > "index". Previous indexes where in a named directory based on slice/shard. > The previous indexes don't appear to have any issues, I just can't > "re-point" the solr cores to them. The Solr.xml file for one of the > servers is: > > <?xml version="1.0" encoding="UTF-8" ?> > <solr persistent="true"> > <cores adminPath="/admin/cores" hostPort="8502"> > <core schema="schema.xml" shard="slice5" instanceDir="test_s5s1/" > name="twitter_s5s1" config="solrconfig.xml" collection="test"/> > </cores> > </solr> > > I think I'm missing exactly what the instanceDir provides. These were the > directories created when I first set up the servers and where the indexes > exist that I want to use. > > Any thought? Or am I just completely off base here in my description of > the issue. > > Chris >
