Can't have 1 replica, even if we can afford ...

to lose the data, because we're unable to produce if the pod that owns that replica goes down. Instead we can wait for 2 acks for important data, and 0 or 1 ack for less important.
Yolean · Nov 3, 2017 · 96641a0 · 96641a0
1 parent bd800d8
commit 96641a0
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/events-kube/topic-ops-kube-events-all-json.yml b/events-kube/topic-ops-kube-events-all-json.yml
@@ -25,8 +25,8 @@ spec:
         - --partitions
         - "1"
         - --replication-factor
-        - "1"
+        - "3"
         - --config
         # 8 days
-        - retention.ms=691200000  
+        - retention.ms=691200000
       restartPolicy: Never