0

Our architecture has 3 nodes in Datacenter: Solr , with SOLR_ENABLED=1

Datacenter: Solr
UN  192.168.8.22  57.02 MB   1          ?       396*****  rack1
UN  192.168.8.23  56.72 MB   1          ?       7365****  rack1
UN  192.168.8.24  59.33 MB   1          ?       92d7****  rack1

We need to integrate Graph to Datastax ,so enabled GRAPH_ENABLED=1 in 192.168.8.22 and our architecture now look like below with 2 DC,Now solr is not working in 192.168.8.22

Datacenter: SearchGraph
UN  192.168.8.22  57.02 MB   1            ?       396***** rack1

.

Datacenter: Solr
UN  192.168.8.23  56.72 MB   1            ?       7365****  rack1
UN  192.168.8.24  59.33 MB   1            ?       92d7****  rack1

1) Can i integrate Solr and graph with in same DC without losing existing data ?

2) Do we need to create separate datacenter node for grap and solr ?

stephen mallette
  • 45,298
  • 5
  • 67
  • 135
Sreeraju V
  • 535
  • 2
  • 5
  • 19

2 Answers2

0

Finally Solved the issue

1) Yes we can integrate Solr and graph on fresh nodes at same DC

Datacenter: SearchGraph

UN 192.168.8.10 116.25 KB 1 ? 9f9909** rack1

UN 192.168.8.11 134.91 KB 1 ? 532344** rack1

UN 192.168.8.12 124.31 KB 1 ? efd6dd** rack1

If data already exist add 

 JVM_OPTS="$JVM_OPTS -Dcassandra.replace_address=**NodeIP**"

to the end of file cassandra-env.sh

2) No,Graph and Solr can be integrate within the same DC. Edit file /etc/default/dse and change as mentioned below

SOLR_ENABLED=1

GRAPH_ENABLED=1
Sreeraju V
  • 535
  • 2
  • 5
  • 19
  • The error you probably noticed in your startup log was related to the Datacenter name changing from "Solr" to "SearchGraph". As a side note for the future, and as a better practice. You would want to decommission the nodes, and then restart them with the appropriate configurations set. This will just ensure you dont experience any data loss. http://docs.datastax.com/en/cassandra/3.0/cassandra/operations/opsDecomissionDC.html – peytoncas Jul 20 '16 at 12:59
0

This is most likely related to the name change of a Datacenter from "Solr" to "SearchGraph" which is what gets assigned in SimpleSnitch configuration. The way you would want to resolve this is by following the decommission process on the appropriate node, enabling graph and restarting the the node. This will help to ensure that you dont experience any data loss when bringing nodes down. You can find the detailed steps in the documentation linked below.

As you answered above, Graph and Solr can be enabled on the same node by editing /etc/default/dse and adjusting the appropriate flags.

http://docs.datastax.com/en/cassandra/3.0/cassandra/operations/opsDecomissionDC.html

peytoncas
  • 755
  • 3
  • 9