Hello All, i have added glue metadata using cli. I...
# getting-started
l
Hello All, i have added glue metadata using cli. I can see the datasets on ui now. But not lineage. Any idea how to add lineage for glue source
c
@lemon-engine-23512 There are few configurations, which can be used to control the lineage. I hope you have taken look at this. Can you share your recipe to confirm if any change is needed over there.
l
Thank you for responding @careful-pilot-86309 . I just used this recipi source: type: glue config: # Coordinates aws_region: "my-aws-region" sink: # sink configs
c
Please set
emit_s3_lineage
to true
l
Will so that and try. The previous recipi has already loaded the metadata. Do i run the recipi again?
@careful-pilot-86309 or should i remove the previous metadata?
c
no need to remove it will get overwritten
l
Thank you @careful-pilot-86309 . One more thing. I have set up datahub quickstart on aws ec2 machine. Usually i stop the instance at the end of day as its not in use. But when i start it again, datahub service is not up. In containers i can only see datahub actions. And ui not accessible. How do i start the services again
c
Youbcan do datahub docker quickstart
Can put it in init script as well
l
@careful-pilot-86309 it worked thanks. Also how i do lineage from glue properties. It holds base tables and pipeline name info, can we build lineage with this using datahub
c
Have you checked this This source actually pulls up lineage automatically based on config parameters.
l
Yeah i did, and used the recipi as mentioned. But it was only able to bring s3 path into lineage. Not rhe base tables or piplines
May be am missing something ? @careful-pilot-86309 i just havr region config and emit_s3_lineage conf in my recipi
I saw there is a 'extract_transform' config. Does this extract only jobs created in glue job or all transformations o a dataset
c
@gentle-hamburger-31302 can you please take it forward with glue details?
l
@gentle-hamburger-31302 would appreciate any help. So i have added glue to my datahub and now i see the datasets and lineage(s3 location). But i want the lineage to display the upstream base tables as well. We have added a custom property in glue with base tables and pipeline name info. Will glue be able to build lineage on this? Or can we do some custom lineage to get this ?
Like glue has config
emit_s3_lineage
. When enabled reads location property and creates lineage for it. can we have a custom property like base tables which reads names of base tables/pipeline in glue property( custom) and creates lineage for that?