Re: migrate(or copy) data from one core1(node2) to anothere core2(node1)

[Mugeesh Husain]

Hello Erick,

I did shutdown all node in solrcloud and copy data directory from non
solcoud, after that i start all of node, but new  data is not refleted.


One doubt 
>> Am i need to shutdown zookeeper node also ? or  also clean
data from zk directory 



--
View this message in context: 
http://lucene.472066.n3.nabble.com/migrate-Copy-data-from-one-core1-server1-to-another-core2-server2-tp4243159p4243582.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: migrate(or copy) data from one core1(node2) to anothere core2(node1)

[Erick Erickson]

Really, this is probably not a great path to go down. If you are saying you
want a leader and follower situation, you should just copy to the leader,
bring it up then use the ADDREPLICA command to add the replica.

That said, I'd seriously consider just defining a new cluster in your SolrCloud
setup the way you want it, then re-index

Best,
Erick

On Fri, Dec 4, 2015 at 6:01 AM, Mugeesh Husain  wrote:
> Hello Erick,
>
> I did shutdown all node in solrcloud and copy data directory from non
> solcoud, after that i start all of node, but new  data is not refleted.
>
>
> One doubt
> >> Am i need to shutdown zookeeper node also ? or  also clean
> data from zk directory
>
>
>
> --
> View this message in context: 
> http://lucene.472066.n3.nabble.com/migrate-Copy-data-from-one-core1-server1-to-another-core2-server2-tp4243159p4243582.html
> Sent from the Solr - User mailing list archive at Nabble.com.

Re: migrate(or copy) data from one core1(node2) to anothere core2(node1)

[Erick Erickson]

Just shut down the SolrCloud instance and copy the index from the
non-cloud to cloud directory. Then bring the cloud instance up. You
should then be fine. This assumes that your SolrCloud instance has
only one shard, which is what I expect given your index size.

After that's done, and assuming you want to add replicas in the
SolrCloud version for HA/DR/Performance reasons, use the ADDREPLICA
Collections API command.

Please try to think about "collections" rather than "cores" when
working with SolrCloud, it'll save grief later on ;)... When you say
you created the core in SolrCloud it worries me. You should use the
Collections API to create a _collection_, NOT fiddling with creating
directories and the like and most certainly not using the _Core_ admin
API. It will be a single-shard collection...

Best,
Erick



On Wed, Dec 2, 2015 at 6:51 AM, Mugeesh Husain  wrote:
> Hello,
>
> I have a 2 solr instance, one is running in solr(non cloud),another one is
> solrcloud mode.
>
> data is indexed in solr mode(non -cloud),now i have creates/define same core
> with same schema in solrcloud instance.
>
> I want to transfer/copy data from one core(non-cloud) to my solrcloud core.
>
> On which way i have to do it.
>
> Note: Index  size is 4.97 GB only.
>
>
>
> --
> View this message in context: 
> http://lucene.472066.n3.nabble.com/migrate-or-copy-data-from-one-core1-node2-to-anothere-core2-node1-tp4243159.html
> Sent from the Solr - User mailing list archive at Nabble.com.

Re: migrate(or copy) data from one core1(node2) to anothere core2(node1)

[Mugeesh Husain]

Thanks Erick,

I am making join operation for multiple core in solrcloud mode.

>>After that's done, and assuming you want to add replicas in the SolrCloud
version for HA/DR/Performance reasons, use the ADDREPLICA  Collections API
command. 

If i split core into shard  then there is any way to use join in multiple
shards.




--
View this message in context: 
http://lucene.472066.n3.nabble.com/migrate-Copy-data-from-one-core1-server1-to-another-core2-server2-tp4243159p4243210.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: migrate(or copy) data from one core1(node2) to anothere core2(node1)

[Erick Erickson]

How are you splitting your core to shards? And why? You should only shard
when you cannot get reasonable performance on a single shard. To increast
the queries-per-second, simply add replicas.

And if at all possible, it would be much less error-prone to just re-index
your data into a collection that was created with shards.

really, this seems like an XY problem. You're asking about specific actions
without explaining what you want to accomplish at a high level.

So why are you going to SolrCloud in the first place? Do you want to go
from a single index to a sharded one? Why? What do you expect to gain?

Best,
Erick

On Wed, Dec 2, 2015 at 9:37 AM, Mugeesh Husain  wrote:
> Thanks Erick,
>
> I am making join operation for multiple core in solrcloud mode.
>
>>>After that's done, and assuming you want to add replicas in the SolrCloud
> version for HA/DR/Performance reasons, use the ADDREPLICA  Collections API
> command.
>
> If i split core into shard  then there is any way to use join in multiple
> shards.
>
>
>
>
> --
> View this message in context: 
> http://lucene.472066.n3.nabble.com/migrate-Copy-data-from-one-core1-server1-to-another-core2-server2-tp4243159p4243210.html
> Sent from the Solr - User mailing list archive at Nabble.com.