1. Janusgraph-Dev
  2. Topics

Pointing Multiple Janusgraph Instances to same HBase giving different results

Jagadeesh Kumar Bhavanasi <jagadeesh...@...>
 

Hi,
We are using Janusgraph 0.2.0, with HBase as storage backend and Solr 5.5 as Indexing engine. Have few questions
  1. Is it possible to create the Janusgraph as a cluster?
  2. As we dont see Clustering feature is not mentioned anywhere, we intended to create multiple instances of Janusgraph servers and pointing to the same HBase instance(to the same Table) along with the same Solr cloud. Is this the right approach?
  3. We are now facing the issues that when we query Janus Server 1 we are getting the different results when we run the same query against another Janus Server.  
Any one help us in this context.

-Regards
Jagadeesh B 
Florian Hockmann <f...@...>
 

Hi Jagadeesh,
many new users of JanusGraph seem to have exactly those questions which is why I recently submitted a pull request to explain how JanusGraph can be deployed like that and that hopefully also reduces this confusion around what a cluster is in the context of JanusGraph. It would be great if you could take a look at the preview version of that doc changes to see if they clear those things up or if anything is still unclear and then give feedback so that we can get that documentation right. You find the pull request itself here and the added chapter that contains different deployment scenarios here. (Feedback is of course welcome from anyone interested in those docs, not only Jagadeesh.)
This should hopefully answer your first two questions already. Hint: The Advanced Scenario should be the right one for you.

Regarding your third question: Do you execute those queries directly after another? Did you just add the data in one query and then try to read it in the other? I never used HBase, so I can't really help you there. My only guess is that it could be caused by caching especially when you just added the data on another JanusGraph Server (that's why I asked about that).

Am Dienstag, 24. Juli 2018 08:42:08 UTC+2 schrieb Jagadeesh Kumar Bhavanasi:

Hi,
We are using Janusgraph 0.2.0, with HBase as storage backend and Solr 5.5 as Indexing engine. Have few questions
  1. Is it possible to create the Janusgraph as a cluster?
  2. As we dont see Clustering feature is not mentioned anywhere, we intended to create multiple instances of Janusgraph servers and pointing to the same HBase instance(to the same Table) along with the same Solr cloud. Is this the right approach?
  3. We are now facing the issues that when we query Janus Server 1 we are getting the different results when we run the same query against another Janus Server.  
Any one help us in this context.

-Regards
Jagadeesh B 
Jason Plurad <plu...@...>
 

For those following along, Florian published a preview of the docs here:

https://florianhockmann.github.io/docs.janusgraph.org/0.2.2-SNAPSHOT/deployment-scenarios.html

It would be great to get more feedback from the community to see if there are any particular gaps or improvements that could make it more clear for new users.


On Tuesday, July 24, 2018 at 10:24:15 AM UTC-4, Florian Hockmann wrote:
Hi Jagadeesh,
many new users of JanusGraph seem to have exactly those questions which is why I recently submitted a pull request to explain how JanusGraph can be deployed like that and that hopefully also reduces this confusion around what a cluster is in the context of JanusGraph. It would be great if you could take a look at the preview version of that doc changes to see if they clear those things up or if anything is still unclear and then give feedback so that we can get that documentation right. You find the pull request itself here and the added chapter that contains different deployment scenarios here. (Feedback is of course welcome from anyone interested in those docs, not only Jagadeesh.)
This should hopefully answer your first two questions already. Hint: The Advanced Scenario should be the right one for you.

Regarding your third question: Do you execute those queries directly after another? Did you just add the data in one query and then try to read it in the other? I never used HBase, so I can't really help you there. My only guess is that it could be caused by caching especially when you just added the data on another JanusGraph Server (that's why I asked about that).

Am Dienstag, 24. Juli 2018 08:42:08 UTC+2 schrieb Jagadeesh Kumar Bhavanasi:
Hi,
We are using Janusgraph 0.2.0, with HBase as storage backend and Solr 5.5 as Indexing engine. Have few questions
  1. Is it possible to create the Janusgraph as a cluster?
  2. As we dont see Clustering feature is not mentioned anywhere, we intended to create multiple instances of Janusgraph servers and pointing to the same HBase instance(to the same Table) along with the same Solr cloud. Is this the right approach?
  3. We are now facing the issues that when we query Janus Server 1 we are getting the different results when we run the same query against another Janus Server.  
Any one help us in this context.

-Regards
Jagadeesh B 

1 - 3 of 3