Unconfigured table exceptions in Janusgraph 0.6.0

Mladen Marović


I wanted to try out the new Janusgraph 0.6.0 release and I encountered some unexpected issues while trying to deploy it on Cassandra 3.11.5.

The first issue occurs during graph creation. I'm trying to set up a ConfiguredGraphFactory. When starting Janusgraph for the first time, it attempts to create its internal tables ('graphindex', 'janusgraph_ids', ...), but sometimes fails to create a table and throws the following exception:

java.lang.RuntimeException: java.lang.reflect.InvocationTargetException
        at org.apache.tinkerpop.gremlin.server.util.ServerGremlinExecutor.<init>(
        at org.apache.tinkerpop.gremlin.server.GremlinServer.<init>(
        at org.apache.tinkerpop.gremlin.server.GremlinServer.<init>(
        at org.janusgraph.graphdb.server.JanusGraphServer.start(
        at org.janusgraph.graphdb.server.JanusGraphServer.main(
Caused by: java.lang.reflect.InvocationTargetException
        at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
        at sun.reflect.NativeConstructorAccessorImpl.newInstance(
        at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(
        at java.lang.reflect.Constructor.newInstance(
        at org.apache.tinkerpop.gremlin.server.util.ServerGremlinExecutor.<init>(
        ... 4 more
Caused by: com.datastax.oss.driver.api.core.servererrors.InvalidQueryException: unconfigured table janusgraph_ids
        at com.datastax.oss.driver.api.core.servererrors.InvalidQueryException.copy(
        at com.datastax.oss.driver.internal.core.util.concurrent.CompletableFutures.getUninterruptibly(
        at com.datastax.oss.driver.internal.core.cql.CqlPrepareSyncProcessor.process(
        at com.datastax.oss.driver.internal.core.cql.CqlPrepareSyncProcessor.process(
        at com.datastax.oss.driver.internal.core.session.DefaultSession.execute(
        at com.datastax.oss.driver.api.core.cql.SyncCqlSession.prepare(
        at org.janusgraph.diskstorage.cql.CQLKeyColumnValueStore.<init>(
        at org.janusgraph.diskstorage.cql.CQLStoreManager.lambda$openDatabase$7(
        at java.util.concurrent.ConcurrentHashMap.computeIfAbsent(
        at org.janusgraph.diskstorage.cql.CQLStoreManager.openDatabase(
        at org.janusgraph.diskstorage.keycolumnvalue.KeyColumnValueStoreManager.openDatabase(
        at org.janusgraph.diskstorage.Backend.initialize(
        at org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.getBackend(
        at org.janusgraph.graphdb.database.StandardJanusGraph.<init>(
        at org.janusgraph.core.JanusGraphFactory.lambda$open$0(
        at java.util.LinkedHashMap.forEach(

Inspecting the keyspaces via cqlsh shows that the keyspace is created, as well as some (but not all) of the tables:

cqlsh> describe keyspaces;

configuration_management_graph  system_auth  system_distributed
system_schema                   system       system_traces     

cqlsh> select table_name from system_schema.tables where keyspace_name = 'configuration_management_graph';


Upon restarting, Janusgraph continues from where it last failed and tries to create the other missing tables. Eventually, cqlsh shows that all tables are created.

cqlsh> select table_name from system_schema.tables where keyspace_name = 'configuration_management_graph';


I tried to test this multiple times by shutting down Janusgraph, dropping its keyspaces in Cassandra, and starting it again. Sometimes it successfully finishes from the first try, sometimes it fails once, and sometimes multiple times, on different tables.

As for the relevant configuration, I've updated it according to the 0.6.0 changelog. I'm using the following gremlin-server.yaml:


port: 8183

evaluationTimeout: 60000



  ConfigurationManagementGraph: /home/test/janusgraph/conf/

      org.janusgraph.graphdb.tinkerpop.plugin.JanusGraphGremlinPlugin: {}
      org.apache.tinkerpop.gremlin.server.jsr223.GremlinServerGremlinPlugin: {}
      org.apache.tinkerpop.gremlin.tinkergraph.jsr223.TinkerGraphGremlinPlugin: {}
          - java.lang.Math
          - java.lang.Math#*
          - /home/test/janusgraph/storage/scripts/globals.groovy

  - className: org.apache.tinkerpop.gremlin.server.op.session.SessionOpProcessor
      sessionTimeout: 28800000
  - className: org.apache.tinkerpop.gremlin.server.op.traversal.TraversalOpProcessor
      cacheExpirationTime: 600000
      cacheMaxSize: 1000

    enabled: true
    interval: 180000

maxInitialLineLength: 4096

maxHeaderSize: 32768

maxChunkSize: 32768

maxContentLength: 2097152

maxAccumulationBufferComponents: 2048

resultIterationBatchSize: 2048

writeBufferLowWaterMark: 131072

writeBufferHighWaterMark: 2097152

threadPoolWorker: 4

gremlinPool: 8

and is as follows:



# disable this property as soon as the upgrade is done to avoid unwanted upgrades

Since this is my first look at Janusgraph 0.6.0, I'm testing it on two local VMs. I'm using Ansible scripts to automate the deployment for Janusgraph, Cassandra and Elasticsearch, and the entire setup has previously been tested on multiple different Janusgraph versions (0.2.0, 0.3.1, 0.4.0, 0.5.1, 0.5.3) and on several different environments. Therefore, I do not expect there to be any issues with Cassandra, Elasticsearch, the deployment scripts, or the setup itself, unless I missed some breaking change in the 0.6.0 changelog that I'm not aware of.

Did anyone else have a similar issue?

Kind regards,

Mladen Marović

P.S. Since this turned out to be a lengthy post, I'll create another post about the other issue I mentioned.

Re: JanusGraph 0.6.0 Binary Driver failing on serialization


Thank you Marc,

I was able to reproduce your sample by console:  :remote connect tinkerpop.server conf/remote.yaml session  # it is GryoMessageSerializerV3d0 - no problem here, but must have serializeResultToString: true 

:remote connect tinkerpop.server conf/remote-graph-binary.yaml session

:> g.V().properties("some")    ==> Server error - Error during serialization: java.lang.IllegalStateException. 

:> g.V().properties("some").hasNext() => true, it is actually working in this context, where server testing the results, not the client after serialization.

It is fine when "remote-graph-binary.yaml" configuration is equipped with serializeResultToString: true. 


That hint should indicate that our server is different that the driver we use in java to remote access, but with more debugging I cannot say where is it... We run JanusGraph services through "bin/", that suppose to run new Binary Driver with gremlin-server-websoc.yaml (+serializeResultToString: true). Then we getting remote access from Java by connection with 

    MessageSerializer<?> serializer = new GraphBinaryMessageSerializerV1(typeSerializerRegistry);

To be sure, we have true for serializeResultToString in the configuration, we even modified that on fly:

    final Map<String, Object> cfg = new HashMap<>();  cfg.put("serializeResultToString",true);     serializer.configure(cfg, null);

With serializeResultToString(true) we have the same exception IllegalStageException as initially described, with (false) it is faling everywhere, on simple command like:

    GraphTraversal<Vertex, Vertex> t = g.V().hasLabel(label).has(property, value);

if (t.hasNext()) {  // java.lang.ClassCastException: java.lang.String cannot be cast to org.apache.tinkerpop.gremlin.process.remote.traversal.RemoteTraverser

so for serializeResultToString(true), all is fine, except "gv().properties()",...  somehow we have good driver config for remote access.

Regards, Chris




Re: Berkeley JE information


Here is how I see this:

- the current public release with direct download is 4 years old 7.5.11, compiled with Java 8 and supporting Java 9 according to its and build script;
- the public Oracle maven repo release is 3 years old 18.3.2, which is according to Oracle a fork from Berkeley JE created specifically for Oracle NoSQL Database, compiled with Java 10;
- the most recent download release of Oracle NoSQL Database includes je version 20.3.17 which is compiled with Java 11 but is not present in public Oracle maven repo (both LATEST and RELEASE return 18.3.12).

Therefore there are 3 different releases of je, where direct download and maven versions are outdated, and Oracle has unofficial plans to replace je with unknown future product.
Too much unsertainty IMHO.

Re: Berkeley JE information


Thanks for this information [I mean the facts, this forum is not intended to spread opinions on third parties].

After, checking, I was surprised myself. The deviation from the JE has occurred since v0.5.0. However, the following file:

Still mentions:
* Oracle BerkeleyDB Java Edition [] (see license below)
This does not seem correct. Can anyone clarify?

Also relevant:

Best wishes,    Marc

Re: JanusGraph 0.6.0 Binary Driver failing on serialization


Re: JanusGraph 0.6.0 Binary Driver failing on serialization


Hi Chris,

Thanks for reporting.  I could confirm the exception you got by using the console and server files from the distribution. I got the right results, though,  when I changed the contents of conf/remote-graph-binary.yaml to:

hosts: [localhost]
port: 8182
serializer: { className: org.apache.tinkerpop.gremlin.driver.ser.GraphBinaryMessageSerializerV1, config: { ioRegistries: [org.janusgraph.graphdb.tinkerpop.JanusGraphIoRegistry], serializeResultToString: true }}

in this way, it matches the configs of gremlin-server in conf/gremlin-server/gremin-server.yaml

I will post a ticket for it.

Best wishes,      Marc

Berkeley JE information


I found the following information that might be important for those who (myself included) use or plan to use Bekeley JE as embedded storage:

"The current public release of BDB JE is 7.5.11. What you are referring to as 18.3.12 is actually from Oracle NoSQL Database Server. This is a fork of BDB JE and has diverged significantly from the 7.5.11 public version. From time to time, we will release updates to the 7.5.x code line but additional public major releases are not planned at this time. While we have not tested internally, it makes sense to me that the 7.5.11 test suite does not run with 18.3.12 (code line has diverged). We do have plans in CY21 to release a new product that will serve as a replacement for BDB JE. Feel free to reach out to me later this spring and I will hopefully have more details (michael.brey@...)"

I looked inside the Oracle NoSQL Database Server 20.3.18, and it includes je version 20.3.17.

IMHO this is so cumbersome in the usual Oracle way, I start looking for a way to get rid of berkeley in my product.

JanusGraph 0.6.0 Binary Driver failing on serialization


We have fully integrated our backend with JS:0.6 as recommended, including DB connection using new driver: GraphBinaryMessageSerializerV1.

All seams o work as before, except:

GraphTraversal<Vertex, ? extends Property<Object>> p = g.V(12345).properties("myproperty");

if (p.hasNext()) {...}


12:02:47.609 [https-jsse-nio-] ERROR o.a.c.c.C.[.[.[.[dispatcherServlet] [175] - Servlet.service() for servlet [dispatcherServlet] in context with path [] threw exception [Request processing failed; nested exception is java.util.concurrent.CompletionException: org.apache.tinkerpop.gremlin.driver.exception.ResponseException: Error during serialization: java.lang.IllegalStateException] with root cause

org.apache.tinkerpop.gremlin.driver.exception.ResponseException: Error during serialization: java.lang.IllegalStateException

at org.apache.tinkerpop.gremlin.driver.Handler$GremlinResponseHandler.channelRead0(

at org.apache.tinkerpop.gremlin.driver.Handler$GremlinResponseHandler.channelRead0(




Above was, and still is fine on p.hasNext() with previously used GryoMessageSerializerV3d0 driver. Workaround we have is to use:

GraphTraversal<Vertex, ? extends Property<Object>> p = g.V(12345).values("myproperty");


if (p.hasNext()) {...}  // now is ok


Why above is failing, is the question....

Regards, Chris 



Re: [DISCUSS] Dropping HBase 1 support


Hi Jan,

When you say "We stopped testing HBase 1 in our CI solution already in branch 0.3" does that mean you are not sure whether HBase1 tests would run on a custom build for 0.4+? If so, the code was already abandoned and dropping it is the only reasonable way to go. Which does not mean such a decision could not be reversed under any circumstances.

Best wishes,   Marc

Re: [DISCUSS] Dropping HBase 1 support

Jansen, Jan

Will the support for BigTable persist when dropping HBase 1? With 0.5, JG uses the HBase 2 as client for Bigtable.

Von: janusgraph-users@... <janusgraph-users@...> im Auftrag von schwartz@... <schwartz@...>
Gesendet: Montag, 13. September 2021 08:05:25
An: janusgraph-users@...
Betreff: Re: [janusgraph-users] [DISCUSS] Dropping HBase 1 support
Will the support for BigTable persist when dropping HBase 1?

We are working with BigTable, and in the days of 0.4 (or was it 0.4.1), BT didn't work, so we had to compile 0.4.x with HBase 1 support.


Re: [DISCUSS] Dropping HBase 1 support


Will the support for BigTable persist when dropping HBase 1?

We are working with BigTable, and in the days of 0.4 (or was it 0.4.1), BT didn't work, so we had to compile 0.4.x with HBase 1 support.


[DISCUSS] Dropping HBase 1 support

Jansen, Jan


I looked into the HBase 1 support after Porunov asked why I want to drop if the builds are passing: We stopped testing HBase 1 in our CI solution already in branch 0.3 that we are happening due to a wrong combination of maven flags.

The Apache HBase already released an alpha version of HBase 3 which would make the support for HBase 1 more complex at the same time.

Any thoughts? Is anyone using HBase 1 with a newer version JG (0.5 or 0.6)? Why couldn't you upgrade to HBase 2?

My Idea would be to drop HBase 1 support. (Currently, HBase 1 support already requires a custom build of JG.)

Greetings, Jan

Re: Performance Improvement

Oleksandr Porunov

Hi Vinayak,

0.6.0 version of JanusGraph is released. I posted some quick tips to improve throughput to your CQL storage here:
I also had a post in LinkedIn with links to relative documentation parts and several better suggestions about internal ExecutorServices usage here:

In 0.6.0 you can improve your CQL throughput drastically using a simple configuration `storage.cql.executor-service.enabled: false` which I definitely recommend to do but you should properly configure throughput related configurations.

Best regards,

Re: Release vs full release?

Oleksandr Porunov

Hi Laura,

Both janusgraph-<version>.zip and janusgraph-full-<version>.zip are the same except that `janusgraph-full-<version>.zip` includes Cassandra and ElasticSearch which can be quickly started with a convenient tool `bin/`. This is the convenient way to test JanusGraph with mixed indices without the need to configure and run your own ElasticSearch and Cassandra. That said, in most cases I would recommend to run your own installation of storage and mixed index backends.
Both these releases have JanusGraph Server (Gremlin Server) included. Thus, in production you should prefer using janusgraph-<version>.zip with your own backend storages installation instead of using janusgraph-full-<version>.zip.

Best regards,

Re: Too low Performance when running PageRank and WCC on Graph500

Oleksandr Porunov

Hi Shipeng,

I didn't check the graph which you refereed but 0.5.3 JanusGraph has some hard limits with Cassandra backend. I would recommend trying 0.6.0 version.
You might want to add some configurations related to your throughput. Something like:
``` ONE
query.batch: true false
# false or true depending on queries
ids.block-size: 1000000
storage.batch-loading: true
storage.cql.local-max-connections-per-host: 5
storage.cql.max-requests-per-connection: 1024
storage.cql.executor-service.enabled: false
storage.parallel-backend-executor-service.core-pool-size: 100

Best regards,

Re: [ANNOUNCE] JanusGraph 0.6.0 Release


Nice!! Gonna give this a spin in a few days

Re: [ANNOUNCE] JanusGraph 0.6.0 Release

Oleksandr Porunov

Re: [ANNOUNCE] JanusGraph 0.6.0 Release

Oleksandr Porunov

Jan Jansen is working on upgrading the docker image to 0.6.0 in this PR:
I believe the Docker image should be available soon.

Re: [ANNOUNCE] JanusGraph 0.6.0 Release


This is great! Was looking forward to this. 

Any ETA for the docker image?

Thanks a lot,

[ANNOUNCE] JanusGraph 0.6.0 Release

Oleksandr Porunov

The JanusGraph Technical Steering Committee is excited to announce the release of JanusGraph 0.6.0.

JanusGraph is an Apache TinkerPop enabled property graph database with support for a variety of storage and indexing backends. Thank you to all of the contributors.

Notable new features in this release include:
  • Upgrade to TinkerPop 3.5.1
  • Java 11 support
  • Spark 3 support
  • Added mixed index usage for count and has("propertyKey") queries
  • Optimized adjacency checks with unique index
  • Index selection algorithms optimization. Added possibility to configure index selection algorithms.
  • Index repair jobs improvements
  • General index construction optimizations
  • Optimized LevenshteinDistance computation used in Fuzzy predicates
  • Update DataStax Cassandra driver to 4.13.0 version
  • Update Lucene / Solr to 8.9.0
  • Metrics collection improvements
  • Many general optimizations in core
  • GraphBinary serialization format support
  • Added new schema maker and improvement of previous schema makers
  • Added DataStax request logger
  • Replaced GremlinServer with JanusGraphServer
  • Added GRPC server to janusgraph-server for basic schema management
  • Transactions improvements
  • Improved inmemory storage backend
  • Added support for Amazon Managed KeySpace
  • Enhanced profiling
  • Added many new configurations to better control storage and index backends
  • Added configuration to use barrier size as batch size limit
  • Added CacheVertex::refresh method to allow clearing vertex cache
  • Added negations to all text predicates
  • Added exists clause to negated Text predicates
  • Make ExecutorService configurable for parallel backend queries and CQL Store Manager
  • Make CQL executor service usage optional
  • Accept optional custom hadoop config in MapReduceIndexManagement
  • Added multi-query and pre-fetch options to transaction builder
  • Added possibility to configure internal Cassandra driver
The release artifacts can be found at this location:

A full binary distribution is provided for user convenience:
A truncated binary distribution is provided:

The online docs can be found here:
To view the resolved issues and commits check the milestone here:

Thank you very much,
Oleksandr Porunov

501 - 520 of 6661