Upgrade ZK and configure autopurge
Nodepool now requires a newer version of ZK, so upgrade to 3.7, the
current stable version.
Also, configure the auto purge cleanup interval (6 hours) -- this will
instruct ZK to remove unused data snapshots every 6 hours in order to
avoid filling the disk.
Both of these changes have already been applied in production as
an emergency fix.
Change-Id: If3b0a9dee5b37c860344f0aa5a577d5b1308ce3b
diff --git a/k8s/zookeeper/values.yaml b/k8s/zookeeper/values.yaml
index 336782b..710672f 100644
--- a/k8s/zookeeper/values.yaml
+++ b/k8s/zookeeper/values.yaml
@@ -17,7 +17,7 @@
## - https://github.com/kubernetes/contrib/blob/master/statefulsets/zookeeper/Makefile#L1
image:
repository: zookeeper # Container image repository for zookeeper container.
- tag: 3.5.5 # Container image tag for zookeeper container.
+ tag: 3.7.1 # Container image tag for zookeeper container.
pullPolicy: IfNotPresent # Image pull criteria for zookeeper container.
service:
@@ -292,7 +292,7 @@
## internal time.
ZK_TICK_TIME: 2000
- ZOO_AUTOPURGE_PURGEINTERVAL: 0
+ ZOO_AUTOPURGE_PURGEINTERVAL: 6
ZOO_AUTOPURGE_SNAPRETAINCOUNT: 3
ZOO_STANDALONE_ENABLED: false
diff --git a/k8s/zookeeper/zookeeper.yaml b/k8s/zookeeper/zookeeper.yaml
index 4b55965..8f39052 100644
--- a/k8s/zookeeper/zookeeper.yaml
+++ b/k8s/zookeeper/zookeeper.yaml
@@ -254,7 +254,7 @@
containers:
- name: zookeeper
- image: "zookeeper:3.5.5"
+ image: "zookeeper:3.7.1"
imagePullPolicy: IfNotPresent
command:
- "/bin/bash"
@@ -306,7 +306,7 @@
- name: ZK_TICK_TIME
value: "2000"
- name: ZOO_AUTOPURGE_PURGEINTERVAL
- value: "0"
+ value: "6"
- name: ZOO_AUTOPURGE_SNAPRETAINCOUNT
value: "3"
- name: ZOO_INIT_LIMIT
