Thanks. I thought it worked like that, but didn't want to jump to conclusions.
On Thu, Jun 4, 2015 at 1:42 PM, Anshum Gupta <ans...@anshumgupta.net> wrote: > Hi Mike, > > Once the SPLITSHARD call completes, it just marks the original shard as > Inactive i.e. it no longer accepts requests. So yes, you would have to use > DELETESHARD ( > > https://cwiki.apache.org/confluence/display/solr/Collections+API#CollectionsAPI-api7 > ) > to clean it up. > > As far as what you see on the admin UI, that information is wrong i.e. the > UI does not respect the state of the shard while displaying them. So, > though the parent shard might be inactive, you still would end up seeing it > as just another active shard. There's an open issue for this one. > > One way to confirm the shard state is by looking at the shard state in > clusterstate.json (or state.json, depending upon the version of Solr you're > using). > > > On Thu, Jun 4, 2015 at 10:35 AM, Mike Thomsen <mikerthom...@gmail.com> > wrote: > > > I thought splitshard was supposed to get rid of the original shard, > > shard1, in this case. Am I missing something? I was expecting the only > two > > remaining shards to be shard1_0 and shard1_1. > > > > The REST call I used was > > > /admin/collections?collection=default-collection&shard=shard1&action=SPLITSHARD > > if that helps. > > > > Attached is a screenshot of the Cloud view in the admin console after > > running splitshard. > > > > Should it look like that? Do I need to delete shard1 now? > > > > Thanks, > > > > Mike > > > > > > -- > Anshum Gupta >
