Hi everyone, We're considering a migration to an HDFS-backed Solr Cloud, both from our 4.2-based Solr Cloud, and a legacy 3.6 classic replication setup. In the end, we hope to unify these two and upgrade to 4.8.1, or 4.9 if that's out in time.
I'm wondering how many of you have experience with migrating to HDFS, and if you managed to do something a little more crafty than a bulk reindex against a new installation. For example, is it possible to do something like join some 4.8, HDFS-backed nodes to your 4.2 setup, add replicas to the new nodes, have things sync over, and then terminate the 4.2 nodes? For our older setup, could I bodge together collections by simply copying the index data into HDFS and building a single shard collection from each one? Would the HDFSDirectoryFactory do OK against an index written using an older codec and on a random access disk? Any information or experiences you might be able to share would be helpful. In the meantime, I'm going to start experimenting with some of these approaches. Thanks! Michael Della Bitta Applications Developer o: +1 646 532 3062 appinions inc. “The Science of Influence Marketing” 18 East 41st Street New York, NY 10017 t: @appinions <https://twitter.com/Appinions> | g+: plus.google.com/appinions <https://plus.google.com/u/0/b/112002776285509593336/112002776285509593336/posts> w: appinions.com <http://www.appinions.com/>