This message was deleted.

Hi Patrik - you will find the metadata for segments (

druid_segments

table) has some different values as you move from HDFS to S3. For example, the

payload

for S3 does not have a

path

field, but includes instead fields for

bucket

,

key

and

S3Schema

- this can require some advanced manipulation

Thank you for your quick reply, Kyle. Yes, I have noticed that, that's what I meant with step 2., update the configuration, so I am mindful of that. My main question is regarding the 1st point, can I just simply copy the segment files and point Druid to them (after setting up the configuration and metadata)?

Yes - you can just sync the segment files. Once you update the metadata, druid knows where to find them and will start to load them onto the historical servers (if they’re not already there)

Thanks a lot!