[GitHub] [airflow] potiuk commented on issue #28603: SFTPOperator does not finish downloading

[GitBox <gi...@apache.org>]

potiuk commented on issue #28603:
URL: https://github.com/apache/airflow/issues/28603#issuecomment-1366733676

    One of the reasons for paramico change was that pysftp did not allow for the kind of configuration that direct paramiko would. You should likely look at `keepalive_interval` and possibly other parameters there. 
   
   The problem that you might experience is because some networking equipment between you and the server could have been restarted (usual thing in the cloud) and if you do not enable keep-alives, you might as a client not even be aware that the other end has stopped responding.
   
   The nature of TCP connection is such that there is no "physical" connection set up - those are just networking packets sent over the network to certain ports. So if you have no "keepalive" enabled, the client might simply not realize the other end is disconnected because if the other end did not have time to signal that it has died or the CLOSE message might not have been delivered. 
   
   Only [TCP keepalives](https://tldp.org/HOWTO/TCP-Keepalive-HOWTO/overview.html) can give you the certainty  that you detect this situation. And they are disabled by default - you need to enable them - in case of SSHook (and SFTPHook derives from it in the new implementation) via the `keepalive_interval` 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@airflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org