Unloading a core is the known way to unregister a solr node in zookeeper (and not use for further querying). It works for me. If you didn't do that like this, unused nodes may remain in the cluster state and Solr may try to use them without a success. I'd suggest to start some machine with the old name, run solr, join the cluster for a while, unload a core to unregister it from the cluster and shutdown host at the end. This way you could have clear cluster state.
On 16 July 2013 14:41, Luis Carlos Guerrero Covo <lcguerreroc...@gmail.com>wrote: > Thanks, I was actually asking about deleting nodes from the cluster state > not cores, unless you can unload cores specific to an already offline node > from zookeeper. > > > On Tue, Jul 16, 2013 at 1:55 AM, Marcin Rzewucki <mrzewu...@gmail.com > >wrote: > > > Hi, > > > > You should use CoreAdmin API (or Solr Admin page) and UNLOAD unneeded > > cores. This will unregister them from the zookeeper (cluster state will > be > > updated), so they won't be used for querying any longer. Solrcloud > restart > > is not needed in this case. > > > > Regards. > > > > > > On 16 July 2013 06:18, Ali, Saqib <docbook....@gmail.com> wrote: > > > > > Hello Luis, > > > > > > I don't think that is possible. If you delete clusterstate.json from > > > zookeeper, you will need to restart the nodes...... I could be very > wrong > > > about this.... > > > > > > Saqib > > > > > > > > > On Mon, Jul 15, 2013 at 8:50 PM, Luis Carlos Guerrero Covo < > > > lcguerreroc...@gmail.com> wrote: > > > > > > > I know that you can clear zookeeper's data directoy using the CLI > with > > > the > > > > clear command, I just want to know if its possible to update the > > > cluster's > > > > state without wiping everything out. Anyone have any > ideas/suggestions? > > > > > > > > > > > > On Mon, Jul 15, 2013 at 11:21 AM, Luis Carlos Guerrero Covo < > > > > lcguerreroc...@gmail.com> wrote: > > > > > > > > > Hi, > > > > > > > > > > Is there an easy way to clear zookeeper of all offline solr nodes > > > without > > > > > restarting the cluster? We are having some stability issues and we > > > think > > > > it > > > > > maybe due to the leader querying old offline nodes. > > > > > > > > > > thank you, > > > > > > > > > > Luis Guerrero > > > > > > > > > > > > > > > > > > > > > -- > > > > Luis Carlos Guerrero Covo > > > > M.S. Computer Engineering > > > > (57) 3183542047 > > > > > > > > > > > > > -- > Luis Carlos Guerrero Covo > M.S. Computer Engineering > (57) 3183542047 >
