Using druid-azure-extensions in index_hadoop task

I have a two node druid cluster and I am loading the

wikiticker example while using azure for deep storage. [druid-azure-extensions]

I dont have any Hadoop cluster configured.

Upon sending the wikiticker-index.json to overlord, an index_hadoop batch ingestion is

launched which fails with the following.

Is there any more config missing? How will druid indexer recognize the wasbs extension??



Are you running with this post-0.12.0 patch?

  • Jon

Yes. I applied the patch. I downloaded

0.12.0 and then replaced the azure extensions jar which I built from master.

It appears that I need Hadoop installed since I can provide the wasbs filesystem impl class in Hadoop’s fs.* configuration.