Memory error when using scan In python to retrieve millions of documents

[liron_gofberg]

I’m using scan in python to retrieve my data
It is working when I ask for thousands of documents but when I ask for millions of records I get an error after a while and that error relate to the memory

What could be happen?

[warkolm (Mark Walkom)]

Please share the full error.
It'd also help if you shared the code that you are using.

[liron_gofberg]

Thanks for answering

scan use scroll
scroll raise the error

I don’t have the code now but it doesn’t matter
I would like you to think about what could happen If we talk about memory error
the error happens when I'm trying to fetch multiple millions of records
but if I ask for 500,000 everything is ok, there is no error

I read about the need to set heap size

If I query with scroll millions of records from Elasticsearch
do I need to set heap size of the master node only? do I need to configure the data nodes too?
do I need to do something else?

the master node need to have available space to assemble the results of all the shards
so if I ask for 20GB of results, The master node need to have 20 GB available?(In heap size?)

I would like to know what memory scroll need to function right with millions of records

Please