1) Depends on your document routing strategy. It sounds like you could be using the compositeId strategy and if so, there's still a hash range assigned to each shard, so you can split the big shards into smaller shards.
2) Since you're replicating in 2 places, when one of your servers crash, there will still be one copy of the shard. When the new server comes online, you should specify -Dshard=ID and the data will be pulled over from the healthy replica using good ol' snapshot replication. btw: ID is the shard Id of the server that crashed 3) No comment on this ;-) Post back if you find a good one. On Tue, Jul 30, 2013 at 2:46 PM, AdityaR <aditya.ravinuth...@gmail.com> wrote: > Hi, > > I have created a solr cloud and have started using it , I have few questions > regarding the set up and it would be really helpful if someone can answer > these. > > Use Case: We have many clients and each clients data is in his own > collection, we currently have 10 server cloud and have distributed clients > on them ( 5 shards for each collection and replicationFactor =2) > > 1) Since we are adding new clients, it will only be matter of time before we > have no space on the 10 servers, when this happens what options do I have? > > 2) Lets say one of the 10 servers crashes and had to be replaced by a > different server, if I add a new server into the cloud will this new server > have the same data as the one that crashed. Meaning does the data get moved > into the new server. > > 3) Are there any monitoring tools for monitoring solr cloud, I have looked > at SPM by sematext and new relic , but am having issues with both the tools. > > Thanks, > Aditya > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Solr-Cloud-Questions-tp4081441.html > Sent from the Solr - User mailing list archive at Nabble.com.
