You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by noman <no...@mtbc.com> on 2020/02/10 12:19:24 UTC

SORLCLOUD

I have created three different solrcloud instance running on three different
ports with external zookeeper 3 instance link with them and when I load the
data in one solrcloud instance, it successfully can be accessible from three
different solrcloud instance.
E.g:
Zookeeper
./zkServer start zoo.cfg (which is running on part 2181)
./zkServer start zoo2.cfg (which is running on part 2182)
./zkServer start zoo3.cfg (which is running on part 2183)

SOLRCloud
solr start -c -p 8983 -z localhost:2181, -z localhost:2181, -z
localhost:2181
solr start -c -p 7574 -z localhost:2182, -z localhost:2182, -z
localhost:2182
solr start -c -p 7575 -z localhost:2183, -z localhost:2183, -z
localhost:2183


These are the starting instance of zookeeper which is link with the
solrcloud instance running with the collection with 2 shard and 2 replicas
which was distributed between these 3 solrcloud instance...
NOW
IF one solrcluod instance gets down then how zookeeper or solrcloud manage
that at run time to fetch the data from other shards or any other solrcloud
instance... please define the changes I have to make and in which
configuration file.. 
Thanks in advance....





--
Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html

Re: SORLCLOUD

Posted by Erick Erickson <er...@gmail.com>.
You’ve misconfigured the startup. Although looking at the 
script help it is a little confusing.

The -z parameter should be the _ensemble_. Pointing each Solr
instance three times at the same ZK instance is not at all what
you need to do.

You should start them up with the “-z” parameter set to something like:

-z zk1:2181,zk2:2181,zk3:2181

where zk1, zk2 and zk3 are three different hosts.

There’s no point at all in running 3 Zookeeper instances on the _same_
physical machine whether or not on different ports. To be robust you
need to be running them on three _different_ machines. If you’re limited
to a single machine, just run one Zookeeper, I do this all the time for 
development purposes.

So to answer your question, when you have properly configured several
Zookeeper instances on different machines, you specify all three
Zookeeper instances in the connection string and all the Solr instances
have access to all the Zookeeper instances. If one node goes down,
Solr can automatically look at the other Zookeeper instances for
the state of the cluster.

I really do suggest you take a bit of time to look over the Zookeeper
introduction, it’ll make a lot more sense then.

Best,
Erick

> On Feb 10, 2020, at 7:19 AM, noman <no...@mtbc.com> wrote:
> 
> I have created three different solrcloud instance running on three different
> ports with external zookeeper 3 instance link with them and when I load the
> data in one solrcloud instance, it successfully can be accessible from three
> different solrcloud instance.
> E.g:
> Zookeeper
> ./zkServer start zoo.cfg (which is running on part 2181) 
> ./zkServer start zoo2.cfg (which is running on part 2182)
> ./zkServer start zoo3.cfg (which is running on part 2183)
> 
> SOLRCloud
> solr start -c -p 8983 -z localhost:2181, -z localhost:2181, -z
> localhost:2181
> solr start -c -p 7574 -z localhost:2182, -z localhost:2182, -z
> localhost:2182
> solr start -c -p 7575 -z localhost:2183, -z localhost:2183, -z
> localhost:2183
> 
> 
> These are the starting instance of zookeeper which is link with the
> solrcloud instance running with the collection with 2 shard and 2 replicas
> which was distributed between these 3 solrcloud instance...
> NOW
> IF one solrcluod instance gets down then how zookeeper or solrcloud manage
> that at run time to fetch the data from other shards or any other solrcloud
> instance... please define the changes I have to make and in which
> configuration file.. 
> Thanks in advance....
> 
> 
> 
> 
> 
> --
> Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html