Correct way of adding new columns in ingestionSpec of Druid dataSource

[Saurabh Pande]

Hi folks,

We had a use-case of adding new columns to our druid dataSource and as per druid docs, druid provides flexibility to add columns to dataSource on the go as schema is maintained at segment level.

We are just confused now how should we add new columns, there two ways as below,

1. Should we directly go on druid coordinator console, open ingestionSpec in data loader, add columns and submit ?

2. Should we first suspend the supervisor, update ingestionSpec as mentioned in above point and then run supervisor again?

We are looking forwards to what are the impacts in both approaches and what should be the recommended way.

Thank you!

Saurabh Pande

[John Kowtko]

Hi Saurabh, 

You should be able to do option #1.   When you edit an existing spec and resubmit, the supervisor job should automatically suspend itself, pick up the new spec and resume operation.  

If you have any further questions or issues with this let us know.

Thanks.  John

[Mrugesh Kadia]

Hi Saurabh,

I have used both methods, and both work fine.

Thanks,

Mrugesh Kadia