[airflow] branch master updated: Change default of `[kubernetes] enable_tcp_keepalive` to `True` (#15338)

[di...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

dimberman pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/airflow.git


The following commit(s) were added to refs/heads/master by this push:
     new 6e31465  Change default of `[kubernetes] enable_tcp_keepalive` to `True` (#15338)
6e31465 is described below

commit 6e31465a30dfd17e2e1409a81600b2e83c910036
Author: Jed Cunningham <66...@users.noreply.github.com>
AuthorDate: Tue Apr 13 08:24:09 2021 -0600

    Change default of `[kubernetes] enable_tcp_keepalive` to `True` (#15338)
    
    We've seen instances of connection resets happening, particularly in
    Azure, that are remedied by enabling tcp_keepalive. Enabling it by
    default should be safe and sane regardless of where we are running.
---
 UPDATING.md                                  |  6 +++++-
 airflow/config_templates/config.yml          |  2 +-
 airflow/config_templates/default_airflow.cfg |  2 +-
 airflow/kubernetes/kube_client.py            | 10 +++++-----
 4 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/UPDATING.md b/UPDATING.md
index 7f059e2..bbacbe0 100644
--- a/UPDATING.md
+++ b/UPDATING.md
@@ -74,7 +74,7 @@ https://developers.google.com/style/inclusive-documentation
 
 Moved the pod launcher from `airflow.kubernetes.pod_launcher` to `airflow.providers.cncf.kubernetes.utils.pod_launcher`
 
-This will alow users to update the pod_launcher for the KubernetesPodOperator without requiring an airflow upgrade
+This will allow users to update the pod_launcher for the KubernetesPodOperator without requiring an airflow upgrade
 
 ### Default `[webserver] worker_refresh_interval` is changed to `6000` seconds
 
@@ -91,6 +91,10 @@ serve as a DagBag cache burst time.
 
 The `default_queue` configuration option has been moved from `[celery]` section to `[operators]` section to allow for re-use between different executors.
 
+### Default `[kubernetes] enable_tcp_keepalive` is changed to `True`
+
+This allows Airflow to work more reliably with some environments (like Azure) by default.
+
 ## Airflow 2.0.1
 
 ### Permission to view Airflow Configurations has been removed from `User` and `Viewer` role
diff --git a/airflow/config_templates/config.yml b/airflow/config_templates/config.yml
index 1e4d4b1..edee338 100644
--- a/airflow/config_templates/config.yml
+++ b/airflow/config_templates/config.yml
@@ -2116,7 +2116,7 @@
       version_added: ~
       type: boolean
       example: ~
-      default: "False"
+      default: "True"
     - name: tcp_keep_idle
       description: |
         When the `enable_tcp_keepalive` option is enabled, TCP probes a connection that has
diff --git a/airflow/config_templates/default_airflow.cfg b/airflow/config_templates/default_airflow.cfg
index c880f3e..c498d46 100644
--- a/airflow/config_templates/default_airflow.cfg
+++ b/airflow/config_templates/default_airflow.cfg
@@ -1045,7 +1045,7 @@ delete_option_kwargs =
 
 # Enables TCP keepalive mechanism. This prevents Kubernetes API requests to hang indefinitely
 # when idle connection is time-outed on services like cloud load balancers or firewalls.
-enable_tcp_keepalive = False
+enable_tcp_keepalive = True
 
 # When the `enable_tcp_keepalive` option is enabled, TCP probes a connection that has
 # been idle for `tcp_keep_idle` seconds.
diff --git a/airflow/kubernetes/kube_client.py b/airflow/kubernetes/kube_client.py
index c83b5e0..aaadbc9 100644
--- a/airflow/kubernetes/kube_client.py
+++ b/airflow/kubernetes/kube_client.py
@@ -85,9 +85,9 @@ def _enable_tcp_keepalive() -> None:
 
     from urllib3.connection import HTTPConnection, HTTPSConnection
 
-    tcp_keep_idle = conf.getint('kubernetes', 'tcp_keep_idle', fallback=120)
-    tcp_keep_intvl = conf.getint('kubernetes', 'tcp_keep_intvl', fallback=30)
-    tcp_keep_cnt = conf.getint('kubernetes', 'tcp_keep_cnt', fallback=6)
+    tcp_keep_idle = conf.getint('kubernetes', 'tcp_keep_idle')
+    tcp_keep_intvl = conf.getint('kubernetes', 'tcp_keep_intvl')
+    tcp_keep_cnt = conf.getint('kubernetes', 'tcp_keep_cnt')
 
     socket_options = [
         (socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1),
@@ -125,10 +125,10 @@ def get_kube_client(
         if config_file is None:
             config_file = conf.get('kubernetes', 'config_file', fallback=None)
 
-    if conf.getboolean('kubernetes', 'enable_tcp_keepalive', fallback=False):
+    if conf.getboolean('kubernetes', 'enable_tcp_keepalive'):
         _enable_tcp_keepalive()
 
-    if not conf.getboolean('kubernetes', 'verify_ssl', fallback=True):
+    if not conf.getboolean('kubernetes', 'verify_ssl'):
         _disable_verify_ssl()
 
     client_conf = _get_kube_config(in_cluster, cluster_context, config_file)