First, why are you reloading the leader? Just as an experiment? I know there's been some JIRA issues with reloading cores and SolrCloud...
Second, what's your evidence that replication didn't happen? For just a few documents the slave index might be updated from the transaction log and you wouldn't necessarily see anything that looked like replication. After you do all this, are your documents visible at all (i.e. in the leader)? Best Erick On Thu, Oct 18, 2012 at 12:52 AM, Minoru Osuka <minoru.os...@gmail.com> wrote: > Hi, > > I am facing replication problem. > I had added a shard replica after the leader's core had been reloaded. I > had expected to start index replication, but it hadn't worked. > Please give me some workaround advice. > > My operation commands are following. > > ZooKeeper localhost:2181 > Solr Shard1 Leader localhost:8983 > Solr Shard1 Replica localhost:7574 > > 1. Start Leader. (localhost:8983) > $ java -Djetty.port=8983 -DzkHost=localhost:2181 > -Dbootstrap_confdir=./solr/collection1/conf > -Dcollection.configName=configuration1 -DnumShards=1 -jar start.jar > > 2. Update the index. (localhost:8983) > $ curl "http://localhost:8983/solr/update/?commit=true"; -H > "Content-Type: text/xml" --data-binary @./example/exampledocs/hd.xml > > 3. Reload Leader. (localhost:8983) > $ curl " > http://localhost:8983/solr/admin/cores?action=RELOAD&core=collection1"; > > 4. Start Replica. (localhost:7574) ## I had expected to start index > replication, but it hadn't worked. > $ java -Djetty.port=7574 -DzkHost=localhost:2181 -jar start.jar > > > Regards, > Minoru > > > -- > --------------------------------------------------- > Minoru Osuka > minoru.os...@gmail.com
