PDH
PDH

Reputation: 76

Unable to start DSE Graph on reset Cassandra

I am testing DSE Graph (using DSE 5.0.7) a on a single node and managed to corrupt it completely. As a result I wiped out all the data files with the intention of rebuilding everything from scratch. On the first restart of Cassandra I forgot to include the -G option but Cassandra came up fine and was viewable from Opscenter, nodetool etc. I shut this down, and cleared out the data directories and restarted Cassandra again, this time with the -G option. It starts up and then shuts itself down with the following warning written to the log:

WARN [main] 2017-06-08 12:59:03,157 NoSpamLogger.java:97 - Failed to create lease HadoopJT.Graph. Possible causes include network/C* issues, the lease being disabled, insufficient replication (you created a new DC and didn't ALTER KEYSPACE dse_leases) and the duration (30000) being different (you have to disable/delete/recreate the lease to change the duration). java.io.IOException: No live replicas for lease HadoopJT.Graph in table dse_leases.leases Nodes [/10.28.98.53] are all down/still starting. at com.datastax.bdp.leasemanager.LeasePlugin.getRemoteMonitor(LeasePlugin.java:538) [dse-core-5.0.7.jar:5.0.7]

After this is the error

ERROR [main] 2017-06-08 12:59:03,182 Configuration.java:2646 - error parsing conf dse-core-default.xml org.xml.sax.SAXParseException: Premature end of file.

with a 0 byte dse-core-default.xml being created. Deleting this and retrying yields the same results so I suspect this is a red herring.

Anyone have any idea how to fix this short of reinstalling everything from scratch?

Upvotes: 1

Views: 121

Answers (1)

PDH
PDH

Reputation: 76

Looks like this might be fixed by removing a very large java_pidnnnnn.hprof file that was sitting in the bin directory. Not sure why this fixed the problem, if anyone has any idea?

Upvotes: 0

Related Questions