Accumulo Store with Kerberos

Hi,

I have an Accumulo cluster (version: 1.7.2-cdh5.5.0) with kerberos enabled and I would like to know how I can configure Gaffer to use my Accumulo cluster as the underlying store.

I have read the doc on Accumulo store and looked at the TheBasics.java from gaffer-doc repo (master branch). I didn’t find any instructions on whether specific settings are required for kerberos.

I downloaded the iterators.jar and distributed it to all my tservers. Then I modified TheBasics.java with my Accumulo properties

        final AccumuloProperties properties = new AccumuloProperties();
        properties.setStoreClass(AccumuloStore.class);
        properties.setInstance("<my_instance>");
        properties.setZookeepers("<my_zk1>:2181,<my_zk2>:2181,<my_zk3>:2181");
        properties.setUser("[e-mail]");
        properties.setPassword("<my_password>");

And I get the following Exception

Exception in thread "main" java.lang.IllegalArgumentException: Could not initialise the store with provided arguments.
	at uk.gov.gchq.gaffer.store.Store.createStore(Store.java:262)
	at uk.gov.gchq.gaffer.graph.Graph$Builder.updateStore(Graph.java:1066)
	at uk.gov.gchq.gaffer.graph.Graph$Builder.build(Graph.java:953)
	at uk.gov.gchq.gaffer.doc.user.walkthrough.TheBasics.run(TheBasics.java:121)
	at uk.gov.gchq.gaffer.doc.user.walkthrough.TheBasics.main(TheBasics.java:161)
Caused by: uk.gov.gchq.gaffer.store.StoreException: Failed to create accumulo connection
	at uk.gov.gchq.gaffer.accumulostore.utils.TableUtils.getConnector(TableUtils.java:219)
	at uk.gov.gchq.gaffer.accumulostore.AccumuloStore.getConnection(AccumuloStore.java:205)
	at uk.gov.gchq.gaffer.accumulostore.utils.TableUtils.ensureTableExists(TableUtils.java:76)
	at uk.gov.gchq.gaffer.accumulostore.AccumuloStore.initialise(AccumuloStore.java:160)
	at uk.gov.gchq.gaffer.store.Store.createStore(Store.java:260)
	... 4 more
Caused by: org.apache.accumulo.core.client.AccumuloException: java.lang.RuntimeException: Cannot use SASL if Hadoop security is not enabled
	at org.apache.accumulo.core.client.impl.ServerClient.execute(ServerClient.java:70)
	at org.apache.accumulo.core.client.impl.ConnectorImpl.<init>(ConnectorImpl.java:73)
	at org.apache.accumulo.core.client.ZooKeeperInstance.getConnector(ZooKeeperInstance.java:308)
	at uk.gov.gchq.gaffer.accumulostore.utils.TableUtils.getConnector(TableUtils.java:217)
	... 8 more
Caused by: java.lang.RuntimeException: Cannot use SASL if Hadoop security is not enabled
	at org.apache.accumulo.core.rpc.SaslConnectionParams.updatePrincipalFromUgi(SaslConnectionParams.java:194)
	at org.apache.accumulo.core.rpc.SaslConnectionParams.<init>(SaslConnectionParams.java:140)
	at org.apache.accumulo.core.client.impl.ClientContext$3.get(ClientContext.java:117)
	at org.apache.accumulo.core.client.impl.ClientContext$3.get(ClientContext.java:109)
	at com.google.common.base.Suppliers$ExpiringMemoizingSupplier.get(Suppliers.java:185)
	at org.apache.accumulo.core.client.impl.ClientContext.getSaslParams(ClientContext.java:182)
	at org.apache.accumulo.core.client.impl.ThriftTransportKey.<init>(ThriftTransportKey.java:44)
	at org.apache.accumulo.core.client.impl.ServerClient.getConnection(ServerClient.java:144)
	at org.apache.accumulo.core.client.impl.ServerClient.getConnection(ServerClient.java:124)
	at org.apache.accumulo.core.client.impl.ServerClient.getConnection(ServerClient.java:119)
	at org.apache.accumulo.core.client.impl.ServerClient.executeRaw(ServerClient.java:100)
	at org.apache.accumulo.core.client.impl.ServerClient.execute(ServerClient.java:64)
	... 11 more

Any suggestions?

[d47853]

Yeah it’s still very high on our priority list. We are hoping to release it along with a number of other version upgrades inc. Hadoop in the early part of next year

[tdong6]

@d47853 Thanks, I disabled SASL and the example code just hangs. I can see it reached zookeeper from the zookeeper log and the connection keeps being started and stopped. After I exited the code, I got

EndOfStreamException: Unable to read additional data from client sessionid 0x17624cb458600f0, likely client has closed socket
        at org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:231)
        at org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)
        at java.lang.Thread.run(Thread.java:748)

which is expected, though I am not sure why it’s stuck there. Then I realised the Gaffer I am using is built on top of Accumulo 1.9.3 and I am using a different version of Accumulo (1.7.2), so that could be the cause. At this point, I feel the best way for me to try Gaffer on Accumulo is probably setting up an Accumulo cluster (1.9.3) with default settings.

I find there is a discussion on upgrading Gaffer to support Accumulo 2.x but don’t see much follow-up if I am not missing anything. Is this still on the roadmap?