whentimeslows/website
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update zookeeper.md (#6565)

Browse Source 
* Update zookeeper.md

As part of a docs sprint in Melbourne I made a very initial pass at trying to tidy this up a little, I would like to have done much much more, but it's a start…

* Updates from feedback

* Format markdown

* FInal tidy
This commit is contained in:
Chris Chinchilla
2018-04-26 23:46:27 +02:00
committed by k8s-ci-robot
parent 89376986cb
commit 691edcfe8d
1 changed files with 253 additions and 339 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/tutorials/stateful-application/zookeeper.md
+163 -249
View File
@@ -7,11 +7,11 @@ reviewers:
- janetkuo
- kow3ns
- smarterclayton
title: Running ZooKeeper, A CP Distributed System
title: Running ZooKeeper, A Distributed System Coordinator
---
{% capture overview %}
This tutorial demonstrates [Apache Zookeeper](https://zookeeper.apache.org) on
This tutorial demonstrates running [Apache Zookeeper](https://zookeeper.apache.org) on
Kubernetes using [StatefulSets](/docs/concepts/workloads/controllers/statefulset/),
[PodDisruptionBudgets](/docs/concepts/workloads/pods/disruptions/#specifying-a-poddisruptionbudget),
and [PodAntiAffinity](/docs/user-guide/node-selection/#inter-pod-affinity-and-anti-affinity-beta-feature).
@@ -22,38 +22,32 @@ and [PodAntiAffinity](/docs/user-guide/node-selection/#inter-pod-affinity-and-an
Before starting this tutorial, you should be familiar with the following
Kubernetes concepts.
* [Pods](/docs/user-guide/pods/single-container/)
* [Cluster DNS](/docs/concepts/services-networking/dns-pod-service/)
* [Headless Services](/docs/concepts/services-networking/service/#headless-services)
* [PersistentVolumes](/docs/concepts/storage/volumes/)
* [PersistentVolume Provisioning](https://github.com/kubernetes/examples/tree/{{page.githubbranch}}/staging/persistent-volume-provisioning/)
* [StatefulSets](/docs/concepts/workloads/controllers/statefulset/)
* [PodDisruptionBudgets](/docs/concepts/workloads/pods/disruptions/#specifying-a-poddisruptionbudget)
* [PodAntiAffinity](/docs/user-guide/node-selection/#inter-pod-affinity-and-anti-affinity-beta-feature)
* [kubectl CLI](/docs/user-guide/kubectl/)
- [Pods](/docs/user-guide/pods/single-container/)
- [Cluster DNS](/docs/concepts/services-networking/dns-pod-service/)
- [Headless Services](/docs/concepts/services-networking/service/#headless-services)
- [PersistentVolumes](/docs/concepts/storage/volumes/)
- [PersistentVolume Provisioning](https://github.com/kubernetes/examples/tree/{{page.githubbranch}}/staging/persistent-volume-provisioning/)
- [StatefulSets](/docs/concepts/workloads/controllers/statefulset/)
- [PodDisruptionBudgets](/docs/concepts/workloads/pods/disruptions/#specifying-a-poddisruptionbudget)
- [PodAntiAffinity](/docs/user-guide/node-selection/#inter-pod-affinity-and-anti-affinity-beta-feature)
- [kubectl CLI](/docs/user-guide/kubectl/)
You will require a cluster with at least four nodes, and each node will require
at least 2 CPUs and 4 GiB of memory. In this tutorial you will cordon and
drain the cluster's nodes. **This means that all Pods on the cluster's nodes
will be terminated and evicted, and the nodes will, temporarily, become
unschedulable.** You should use a dedicated cluster for this tutorial, or you
should ensure that the disruption you cause will not interfere with other
tenants.
You will require a cluster with at least four nodes, and each node requires at least 2 CPUs and 4 GiB of memory. In this tutorial you will cordon and drain the cluster's nodes. **This means that the cluster will terminate and evict all Pods on its nodes, and the nodes will temporarily become unschedulable.** You should use a dedicated cluster for this tutorial, or you should ensure that the disruption you cause will not interfere with other tenants.
This tutorial assumes that your cluster is configured to dynamically provision
This tutorial assumes that you have configured your cluster to dynamically provision
PersistentVolumes. If your cluster is not configured to do so, you
will have to manually provision three 20 GiB volumes prior to starting this
will have to manually provision three 20 GiB volumes before starting this
tutorial.
{% endcapture %}
{% capture objectives %}
After this tutorial, you will know the following.
* How to deploy a ZooKeeper ensemble using StatefulSet.
* How to consistently configure the ensemble using ConfigMaps.
* How to spread the deployment of ZooKeeper servers in the ensemble.
* How to use PodDisruptionBudgets to ensure service availability during planned maintenance.
{% endcapture %}
- How to deploy a ZooKeeper ensemble using StatefulSet.
- How to consistently configure the ensemble using ConfigMaps.
- How to spread the deployment of ZooKeeper servers in the ensemble.
- How to use PodDisruptionBudgets to ensure service availability during planned maintenance.
{% endcapture %}
{% capture lessoncontent %}
@@ -68,22 +62,9 @@ are atomic and sequentially consistent. ZooKeeper ensures this by using the
[Zab](https://pdfs.semanticscholar.org/b02c/6b00bd5dbdbd951fddb00b906c82fa80f0b3.pdf)
consensus protocol to replicate a state machine across all servers in the ensemble.
The ensemble uses the Zab protocol to elect a leader, and
data can not be written until a leader is elected. Once a leader is
elected, the ensemble uses Zab to ensure that all writes are replicated to a
quorum before they are acknowledged and made visible to clients. Without respect
to weighted quorums, a quorum is a majority component of the ensemble containing
the current leader. For instance, if the ensemble has three servers, a component
that contains the leader and one other server constitutes a quorum. If the
ensemble can not achieve a quorum, data can not be written.
The ensemble uses the Zab protocol to elect a leader, and the ensemble cannot write data until that election is complete. Once complete, the ensemble uses Zab to ensure that it replicates all writes to a quorum before it acknowledges and makes them visible to clients. Without respect to weighted quorums, a quorum is a majority component of the ensemble containing the current leader. For instance, if the ensemble has three servers, a component that contains the leader and one other server constitutes a quorum. If the ensemble can not achieve a quorum, the ensemble cannot write data.
ZooKeeper servers keep their entire state machine in memory, but every mutation
is written to a durable WAL (Write Ahead Log) on storage media. When a server
crashes, it can recover its previous state by replaying the WAL. In order to
prevent the WAL from growing without bound, ZooKeeper servers will periodically
snapshot their in memory state to storage media. These snapshots can be loaded
directly into memory, and all WAL entries that preceded the snapshot may be
safely discarded.
ZooKeeper servers keep their entire state machine in memory, and write every mutation to a durable WAL (Write Ahead Log) on storage media. When a server crashes, it can recover its previous state by replaying the WAL. To prevent the WAL from growing without bound, ZooKeeper servers will periodically snapshot their in memory state to storage media. These snapshots can be loaded directly into memory, and all WAL entries that preceded the snapshot may be discarded.
## Creating a ZooKeeper Ensemble
@@ -95,12 +76,12 @@ and a [StatefulSet](/docs/concepts/workloads/controllers/statefulset/).
{% include code.html language="yaml" file="zookeeper.yaml" ghlink="/docs/tutorials/stateful-application/zookeeper.yaml" %}
Open a command terminal, and use
[`kubectl apply`](/docs/user-guide/kubectl/{{page.version}}/#apply) to create the
Open a terminal, and use the
[`kubectl apply`](/docs/user-guide/kubectl/{{page.version}}/#apply) command to create the
manifest.
```shell
kubectl apply -f https://raw.githubusercontent.com/kubernetes/website/master/docs/tutorials/stateful-application/zookeeper.yaml
kubectl apply -f https://k8s.io/docs/tutorials/stateful-application/zookeeper.yaml
```
This creates the `zk-hs` Headless Service, the `zk-cs` Service,
@@ -146,11 +127,7 @@ a [ZooKeeper](http://www-us.apache.org/dist/zookeeper/stable/) server.
### Facilitating Leader Election
As there is no terminating algorithm for electing a leader in an anonymous
network, Zab requires explicit membership configuration in order to perform
leader election. Each server in the ensemble needs to have a unique
identifier, all servers need to know the global set of identifiers, and each
identifier needs to be associated with a network address.
Because there is no terminating algorithm for electing a leader in an anonymous network, Zab requires explicit membership configuration to perform leader election. Each server in the ensemble needs to have a unique identifier, all servers need to know the global set of identifiers, and each identifier needs to be associated with a network address.
Use [`kubectl exec`](/docs/user-guide/kubectl/{{page.version}}/#exec) to get the hostnames
of the Pods in the `zk` StatefulSet.
@@ -159,10 +136,7 @@ of the Pods in the `zk` StatefulSet.
for i in 0 1 2; do kubectl exec zk-$i -- hostname; done
```
The StatefulSet controller provides each Pod with a unique hostname based on its
ordinal index. The hostnames take the form `<statefulset name>-<ordinal index>`.
As the `replicas` field of the `zk` StatefulSet is set to `3`, the Set's
controller creates three Pods with their hostnames set to `zk-0`, `zk-1`, and
The StatefulSet controller provides each Pod with a unique hostname based on its ordinal index. The hostnames take the form of `<statefulset name>-<ordinal index>`. Because the `replicas` field of the `zk` StatefulSet is set to `3`, the Set's controller creates three Pods with their hostnames set to `zk-0`, `zk-1`, and
`zk-2`.
```shell
@@ -171,18 +145,15 @@ zk-1
zk-2
```
The servers in a ZooKeeper ensemble use natural numbers as unique identifiers, and
each server's identifier is stored in a file called `myid` in the server's
data directory.
The servers in a ZooKeeper ensemble use natural numbers as unique identifiers, and store each server's identifier in a file called `myid` in the server's data directory.
Examine the contents of the `myid` file for each server.
To examine the contents of the `myid` file for each server use the following command.
```shell
for i in 0 1 2; do echo "myid zk-$i";kubectl exec zk-$i -- cat /var/lib/zookeeper/data/myid; done
```
As the identifiers are natural numbers and the ordinal indices are non-negative
integers, you can generate an identifier by adding one to the ordinal.
Because the identifiers are natural numbers and the ordinal indices are non-negative integers, you can generate an identifier by adding 1 to the ordinal.
```shell
myid zk-0
@@ -193,7 +164,7 @@ myid zk-2
3
```
Get the FQDN (Fully Qualified Domain Name) of each Pod in the `zk` StatefulSet.
To get the Fully Qualified Domain Name (FQDN) of each Pod in the `zk` StatefulSet use the following command.
```shell
for i in 0 1 2; do kubectl exec zk-$i -- hostname -f; done
@@ -208,18 +179,15 @@ zk-1.zk-hs.default.svc.cluster.local
zk-2.zk-hs.default.svc.cluster.local
```
The A records in [Kubernetes DNS](/docs/concepts/services-networking/dns-pod-service/) resolve the FQDNs to the Pods' IP addresses.
If the Pods are rescheduled, the A records will be updated with the Pods' new IP
addresses, but the A record's names will not change.
The A records in [Kubernetes DNS](/docs/concepts/services-networking/dns-pod-service/) resolve the FQDNs to the Pods' IP addresses. If Kubernetes reschedules the Pods, it will update the A records with the Pods' new IP addresses, but the A records names will not change.
ZooKeeper stores its application configuration in a file named `zoo.cfg`. Use
`kubectl exec` to view the contents of the `zoo.cfg` file in the `zk-0` Pod.
ZooKeeper stores its application configuration in a file named `zoo.cfg`. Use `kubectl exec` to view the contents of the `zoo.cfg` file in the `zk-0` Pod.
```
```shell
kubectl exec zk-0 -- cat /opt/zookeeper/conf/zoo.cfg
```
For the `server.1`, `server.2`, and `server.3` properties at the bottom of
In the `server.1`, `server.2`, and `server.3` properties at the bottom of
the file, the `1`, `2`, and `3` correspond to the identifiers in the
ZooKeeper servers' `myid` files. They are set to the FQDNs for the Pods in
the `zk` StatefulSet.
@@ -243,15 +211,11 @@ server.3=zk-2.zk-hs.default.svc.cluster.local:2888:3888
### Achieving Consensus
Consensus protocols require that the identifiers of each participant be
unique. No two participants in the Zab protocol should claim the same unique
identifier. This is necessary to allow the processes in the system to agree on
which processes have committed which data. If two Pods were launched with the
same ordinal, two ZooKeeper servers would both identify themselves as the same
server.
Consensus protocols require that the identifiers of each participant be unique. No two participants in the Zab protocol should claim the same unique identifier. This is necessary to allow the processes in the system to agree on which processes have committed which data. If two Pods are launched with the same ordinal, two ZooKeeper servers would both identify themselves as the same server.
```shell
kubectl get pods -w -l app=zk
NAME READY STATUS RESTARTS AGE
zk-0 0/1 Pending 0 0s
zk-0 0/1 Pending 0 0s
@@ -270,8 +234,8 @@ zk-2 0/1 Running 0 19s
zk-2 1/1 Running 0 40s
```
The A records for each Pod are only entered when the Pod becomes Ready. Therefore,
the FQDNs of the ZooKeeper servers will only resolve to a single endpoint, and that
The A records for each Pod are entered when the Pod becomes Ready. Therefore,
the FQDNs of the ZooKeeper servers will resolve to a single endpoint, and that
endpoint will be the unique ZooKeeper server claiming the identity configured
in its `myid` file.
@@ -290,39 +254,31 @@ server.2=zk-1.zk-hs.default.svc.cluster.local:2888:3888
server.3=zk-2.zk-hs.default.svc.cluster.local:2888:3888
```
When the servers use the Zab protocol to attempt to commit a value, they will
either achieve consensus and commit the value (if leader election has succeeded
and at least two of the Pods are Running and Ready), or they will fail to do so
(if either of the aforementioned conditions are not met). No state will arise
where one server acknowledges a write on behalf of another.
When the servers use the Zab protocol to attempt to commit a value, they will either achieve consensus and commit the value (if leader election has succeeded and at least two of the Pods are Running and Ready), or they will fail to do so (if either of the conditions are not met). No state will arise where one server acknowledges a write on behalf of another.
### Sanity Testing the Ensemble
The most basic sanity test is to write some data to one ZooKeeper server and
The most basic sanity test is to write data to one ZooKeeper server and
to read the data from another.
Use the `zkCli.sh` script to write `world` to the path `/hello` on the `zk-0` Pod.
The command below executes the `zkCli.sh` script to write `world` to the path `/hello` on the `zk-0` Pod in the ensemble.
```shell
kubectl exec zk-0 zkCli.sh create /hello world
```
This will write `world` to the `/hello` path in the ensemble.
```shell
WATCHER::
WatchedEvent state:SyncConnected type:None path:null
Created /hello
```
Get the data from the `zk-1` Pod.
To get the data from the `zk-1` Pod use the following command.
```shell
kubectl exec zk-1 zkCli.sh get /hello
```
The data that you created on `zk-0` is available on all of the servers in the
The data that you created on `zk-0` is available on all the servers in the
ensemble.
```shell
@@ -349,9 +305,9 @@ As mentioned in the [ZooKeeper Basics](#zookeeper-basics) section,
ZooKeeper commits all entries to a durable WAL, and periodically writes snapshots
in memory state, to storage media. Using WALs to provide durability is a common
technique for applications that use consensus protocols to achieve a replicated
state machine and for storage applications in general.
state machine.
Use [`kubectl delete`](/docs/user-guide/kubectl/{{page.version}}/#delete) to delete the
Use the [`kubectl delete`](/docs/user-guide/kubectl/{{page.version}}/#delete) command to delete the
`zk` StatefulSet.
```shell
@@ -381,14 +337,14 @@ zk-0 0/1 Terminating 0 11m
zk-0 0/1 Terminating 0 11m
zk-0 0/1 Terminating 0 11m
```
Reapply the manifest in `zookeeper.yaml`.
```shell
kubectl apply -f https://raw.githubusercontent.com/kubernetes/website/master/docs/tutorials/stateful-application/zookeeper.yaml
kubectl apply -f https://k8s.io/docs/tutorials/stateful-application/zookeeper.yaml
```
The `zk` StatefulSet will be created, but, as they already exist, the other API
Objects in the manifest will not be modified.
This creates the `zk` StatefulSet object, but the other API objects in the manifest are not modified because they already exist.
Watch the StatefulSet controller recreate the StatefulSet's Pods.
@@ -417,15 +373,14 @@ zk-2 0/1 Running 0 19s
zk-2 1/1 Running 0 40s
```
Get the value you entered during the [sanity test](#sanity-testing-the-ensemble),
Use the command below to get the value you entered during the [sanity test](#sanity-testing-the-ensemble),
from the `zk-2` Pod.
```shell
kubectl exec zk-2 zkCli.sh get /hello
```
Even though all of the Pods in the `zk` StatefulSet have been terminated and
recreated, the ensemble still serves the original value.
Even though you terminated and recreated all of the Pods in the `zk` StatefulSet, the ensemble still serves the original value.
```shell
WATCHER::
@@ -445,8 +400,7 @@ dataLength = 5
numChildren = 0
```
The `volumeClaimTemplates` field, of the `zk` StatefulSet's `spec`, specifies a
PersistentVolume that will be provisioned for each Pod.
The `volumeClaimTemplates` field of the `zk` StatefulSet's `spec` specifies a PersistentVolume provisioned for each Pod.
```yaml
volumeClaimTemplates:
@@ -461,18 +415,16 @@ volumeClaimTemplates:
storage: 20Gi
```
The `StatefulSet` controller generates a `PersistentVolumeClaim` for each Pod in
the `StatefulSet`.
The StatefulSet controller generates a PersistentVolumeClaim for each Pod in
the StatefulSet.
Get the StatefulSet's PersistentVolumeClaims.
Use the following command to get the `StatefulSet`'s `PersistentVolumeClaims`.
```shell
kubectl get pvc -l app=zk
```
When the StatefulSet recreated its Pods, the Pods' PersistentVolumes were
remounted.
When the `StatefulSet` recreated its Pods, it remounts the Pods' PersistentVolumes.
```shell
NAME STATUS VOLUME CAPACITY ACCESSMODES AGE
@@ -481,8 +433,7 @@ datadir-zk-1 Bound pvc-bedd27d2-bcb1-11e6-994f-42010a800002 20Gi R
datadir-zk-2 Bound pvc-bee0817e-bcb1-11e6-994f-42010a800002 20Gi RWO 1h
```
The `volumeMounts` section of the StatefulSet's container `template` causes the
PersistentVolumes to be mounted to the ZooKeeper servers' data directories.
The `volumeMounts` section of the `StatefulSet`'s container `template` mounts the PersistentVolumes in the ZooKeeper servers' data directories.
```shell
volumeMounts:
@@ -490,16 +441,16 @@ volumeMounts:
mountPath: /var/lib/zookeeper
```
When a Pod in the `zk` StatefulSet is (re)scheduled, it will always have the
same PersistentVolume mounted to the ZooKeeper server's data directory.
Even when the Pods are rescheduled, all of the writes made to the ZooKeeper
servers' WALs, and all of their snapshots, remain durable.
When a Pod in the `zk` `StatefulSet` is (re)scheduled, it will always have the
same `PersistentVolume` mounted to the ZooKeeper server's data directory.
Even when the Pods are rescheduled, all the writes made to the ZooKeeper
servers' WALs, and all their snapshots, remain durable.
## Ensuring Consistent Configuration
As noted in the [Facilitating Leader Election](#facilitating-leader-election) and
[Achieving Consensus](#achieving-consensus) sections, the servers in a
ZooKeeper ensemble require consistent configuration in order to elect a leader
ZooKeeper ensemble require consistent configuration to elect a leader
and form a quorum. They also require consistent configuration of the Zab protocol
in order for the protocol to work correctly over a network. In our example we
achieve consistent configuration by embedding the configuration directly into
@@ -507,10 +458,10 @@ the manifest.
Get the `zk` StatefulSet.
```shell{% raw %}
kubectl get sts zk -o yaml
...
command:
```shell
kubectl get sts zk -o yaml
…
command:
- sh
- -c
- "start-zookeeper \
@@ -531,19 +482,18 @@ Get the `zk` StatefulSet.
--max_session_timeout=40000 \
--min_session_timeout=4000 \
--log_level=INFO"
...
```{% endraw %}
…
```
Notice that the command used to start the ZooKeeper servers passed the configuration
as command line parameter. Environment variables are another, equally good, way to
pass configuration to ensemble.
The command used to start the ZooKeeper servers passed the configuration as command line parameter. You can also use environment variables to pass configuration to the ensemble.
### Configuring Logging
One of the files generated by the `zkGenConfig.sh` script controls ZooKeeper's logging.
ZooKeeper uses [Log4j](http://logging.apache.org/log4j/2.x/), and, by default,
it uses a time and size based rolling file appender for its logging configuration.
Get the logging configuration from one of Pods in the `zk` StatefulSet.
Use the command below to get the logging configuration from one of Pods in the `zk` `StatefulSet`.
```shell
kubectl exec zk-0 cat /usr/etc/zookeeper/log4j.properties
@@ -562,21 +512,15 @@ log4j.appender.CONSOLE.layout=org.apache.log4j.PatternLayout
log4j.appender.CONSOLE.layout.ConversionPattern=%d{ISO8601} [myid:%X{myid}] - %-5p [%t:%C{1}@%L] - %m%n
```
This is the simplest possible way to safely log inside the container. As the
application's logs are being written to standard out, Kubernetes will handle
log rotation for you. Kubernetes also implements a sane retention policy that
ensures application logs written to standard out and standard error do not
exhaust local storage media.
This is the simplest possible way to safely log inside the container. Because the applications write logs to standard out, Kubernetes will handle log rotation for you. Kubernetes also implements a sane retention policy that ensures application logs written to standard out and standard error do not exhaust local storage media.
Use [`kubectl logs`](/docs/user-guide/kubectl/{{page.version}}/#logs) to retrieve the last
few log lines from one of the Pods.
Use [`kubectl logs`](/docs/user-guide/kubectl/{{page.version}}/#logs) to retrieve the last 20 log lines from one of the Pods.
```shell
kubectl logs zk-0 --tail 20
```
Application logs that are written to standard out or standard error are viewable
using `kubectl logs` and from the Kubernetes Dashboard.
You can view application logs written to standard out or standard error using `kubectl logs` and from the Kubernetes Dashboard.
```shell
2016-12-06 19:34:16,236 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52740
@@ -601,22 +545,21 @@ using `kubectl logs` and from the Kubernetes Dashboard.
2016-12-06 19:34:46,230 [myid:1] - INFO [Thread-1142:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52768 (no session established for client)
```
Kubernetes also supports more powerful, but more complex, logging integrations
with [Logging Using Stackdriver](/docs/tasks/debug-application-cluster/logging-stackdriver/)
and [Logging Using Elasticsearch and Kibana](/docs/tasks/debug-application-cluster/logging-elasticsearch-kibana/).
For cluster level log shipping and aggregation, you should consider deploying a
[sidecar](http://blog.kubernetes.io/2015/06/the-distributed-system-toolkit-patterns.html)
Kubernetes supports more powerful, but more complex, logging integrations
with [Stackdriver](/docs/tasks/debug-application-cluster/logging-stackdriver/)
and [Elasticsearch and Kibana](/docs/tasks/debug-application-cluster/logging-elasticsearch-kibana/).
For cluster level log shipping and aggregation, consider deploying a [sidecar](http://blog.kubernetes.io/2015/06/the-distributed-system-toolkit-patterns.html)
container to rotate and ship your logs.
### Configuring a Non-Privileged User
The best practices with respect to allowing an application to run as a privileged
The best practices to allow an application to run as a privileged
user inside of a container are a matter of debate. If your organization requires
that applications be run as a non-privileged user you can use a
that applications run as a non-privileged user you can use a
[SecurityContext](/docs/tasks/configure-pod-container/security-context/) to control the user that
the entry point runs as.
The `zk` StatefulSet's Pod `template` contains a SecurityContext.
The `zk` `StatefulSet`'s Pod `template` contains a `SecurityContext`.
```yaml
securityContext:
@@ -642,19 +585,15 @@ F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD
0 S zookeep+ 27 1 0 80 0 - 1155556 - 20:46 ? 00:00:19 /usr/lib/jvm/java-8-openjdk-amd64/bin/java -Dzookeeper.log.dir=/var/log/zookeeper -Dzookeeper.root.logger=INFO,CONSOLE -cp /usr/bin/../build/classes:/usr/bin/../build/lib/*.jar:/usr/bin/../share/zookeeper/zookeeper-3.4.9.jar:/usr/bin/../share/zookeeper/slf4j-log4j12-1.6.1.jar:/usr/bin/../share/zookeeper/slf4j-api-1.6.1.jar:/usr/bin/../share/zookeeper/netty-3.10.5.Final.jar:/usr/bin/../share/zookeeper/log4j-1.2.16.jar:/usr/bin/../share/zookeeper/jline-0.9.94.jar:/usr/bin/../src/java/lib/*.jar:/usr/bin/../etc/zookeeper: -Xmx2G -Xms2G -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.local.only=false org.apache.zookeeper.server.quorum.QuorumPeerMain /usr/bin/../etc/zookeeper/zoo.cfg
```
By default, when the Pod's PersistentVolume is mounted to the ZooKeeper server's
data directory, it is only accessible by the root user. This configuration
prevents the ZooKeeper process from writing to its WAL and storing its snapshots.
By default, when the Pod's PersistentVolumes is mounted to the ZooKeeper server's data directory, it is only accessible by the root user. This configuration prevents the ZooKeeper process from writing to its WAL and storing its snapshots.
Get the file permissions of the ZooKeeper data directory on the `zk-0` Pod.
Use the command below to get the file permissions of the ZooKeeper data directory on the `zk-0` Pod.
```shell
kubectl exec -ti zk-0 -- ls -ld /var/lib/zookeeper/data
```
As the `fsGroup` field of the `securityContext` object is set to 1000,
the ownership of the Pods' PersistentVolumes is set to the zookeeper group,
and the ZooKeeper process is able to successfully read and write its data.
Because the `fsGroup` field of the `securityContext` object is set to 1000, the ownership of the Pods' PersistentVolumes is set to the zookeeper group, and the ZooKeeper process is able to read and write its data.
```shell
drwxr-sr-x 3 zookeeper zookeeper 4096 Dec 5 20:45 /var/lib/zookeeper/data
@@ -663,7 +602,7 @@ drwxr-sr-x 3 zookeeper zookeeper 4096 Dec 5 20:45 /var/lib/zookeeper/data
## Managing the ZooKeeper Process
The [ZooKeeper documentation](https://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_supervision)
indicates that "You will want to have a supervisory process that
mentions that "You will want to have a supervisory process that
manages each of your ZooKeeper server processes (JVM)." Utilizing a watchdog
(supervisory process) to restart failed processes in a distributed system is a
common pattern. When deploying an application in Kubernetes, rather than using
@@ -672,20 +611,21 @@ watchdog for your application.
### Updating the Ensemble
The `zk` StatefulSet is configured to use the RollingUpdate update strategy.
The `zk` `StatefulSet` is configured to use the `RollingUpdate` update strategy.
You can use `kubectl patch` to update the number of `cpus` allocated to the servers.
```shell{% raw %}
```shell
kubectl patch sts zk --type='json' -p='[{"op": "replace", "path": "/spec/template/spec/containers/0/resources/requests/cpu", "value":"0.3"}]'
statefulset "zk" patched
```{% endraw %}
```
Use `kubectl rollout status` to watch the status of the update.
```shell
kubectl rollout status sts/zk
waiting for statefulset rolling update to complete 0 pods at revision zk-5db4499664...
Waiting for 1 pods to be ready...
Waiting for 1 pods to be ready...
@@ -698,37 +638,36 @@ Waiting for 1 pods to be ready...
statefulset rolling update complete 3 pods at revision zk-5db4499664...
```
The Pods are terminated, one at a time, in reverse ordinal order, and they
are recreated with the new configuration. This ensures that quorum is maintained
during a rolling update.
This terminates the Pods, one at a time, in reverse ordinal order, and recreates them with the new configuration. This ensures that quorum is maintained during a rolling update.
Use `kubectl rollout history` to view a history or previous configurations.
Use the `kubectl rollout history` command to view a history or previous configurations.
```shell
kubectl rollout history sts/zk
statefulsets "zk"
REVISION
1
2
```
Use `kubectl rollout undo` to roll back the modification.
Use the `kubectl rollout undo` command to roll back the modification.
```shell
kubectl rollout undo sts/zk
statefulset "zk" rolled back
```
### Handling Process Failure
[Restart Policies](/docs/user-guide/pod-states/#restartpolicy) control how
Kubernetes handles process failures for the entry point of the container in a Pod.
For Pods in a StatefulSet, the only appropriate RestartPolicy is Always, and this
For Pods in a `StatefulSet`, the only appropriate `RestartPolicy` is Always, and this
is the default value. For stateful applications you should **never** override
the default policy.
Examine the process tree for the ZooKeeper server running in the `zk-0` Pod.
Use the following command to examine the process tree for the ZooKeeper server running in the `zk-0` Pod.
```shell
kubectl exec zk-0 -- ps -ef
@@ -737,31 +676,25 @@ kubectl exec zk-0 -- ps -ef
The command used as the container's entry point has PID 1, and
the ZooKeeper process, a child of the entry point, has PID 23.
```
```shell
UID PID PPID C STIME TTY TIME CMD
zookeep+ 1 0 0 15:03 ? 00:00:00 sh -c zkGenConfig.sh && zkServer.sh start-foreground
zookeep+ 27 1 0 15:03 ? 00:00:03 /usr/lib/jvm/java-8-openjdk-amd64/bin/java -Dzookeeper.log.dir=/var/log/zookeeper -Dzookeeper.root.logger=INFO,CONSOLE -cp /usr/bin/../build/classes:/usr/bin/../build/lib/*.jar:/usr/bin/../share/zookeeper/zookeeper-3.4.9.jar:/usr/bin/../share/zookeeper/slf4j-log4j12-1.6.1.jar:/usr/bin/../share/zookeeper/slf4j-api-1.6.1.jar:/usr/bin/../share/zookeeper/netty-3.10.5.Final.jar:/usr/bin/../share/zookeeper/log4j-1.2.16.jar:/usr/bin/../share/zookeeper/jline-0.9.94.jar:/usr/bin/../src/java/lib/*.jar:/usr/bin/../etc/zookeeper: -Xmx2G -Xms2G -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.local.only=false org.apache.zookeeper.server.quorum.QuorumPeerMain /usr/bin/../etc/zookeeper/zoo.cfg
```
In one terminal watch the Pods in the `zk` StatefulSet.
In another terminal watch the Pods in the `zk` `StatefulSet` with the following command.
```shell
kubectl get pod -w -l app=zk
```
In another terminal, kill the ZooKeeper process in Pod `zk-0`.
In another terminal, terminate the ZooKeeper process in Pod `zk-0` with the following command.
```shell
kubectl exec zk-0 -- pkill java
kubectl exec zk-0 -- pkill java
```
The death of the ZooKeeper process caused its parent process to terminate. As
the RestartPolicy of the container is Always, the parent process was relaunched.
The termination of the ZooKeeper process caused its parent process to terminate. Because the `RestartPolicy` of the container is Always, it restarted the parent process.
```shell
NAME READY STATUS RESTARTS AGE
@@ -774,25 +707,21 @@ zk-0 0/1 Running 1 29m
zk-0 1/1 Running 1 29m
```
If your application uses a script (such as zkServer.sh) to launch the process
If your application uses a script (such as `zkServer.sh`) to launch the process
that implements the application's business logic, the script must terminate with the
child process. This ensures that Kubernetes will restart the application's
container when the process implementing the application's business logic fails.
### Testing for Liveness
Configuring your application to restart failed processes is not sufficient to
keep a distributed system healthy. There are many scenarios where
Configuring your application to restart failed processes is not enough to
keep a distributed system healthy. There are scenarios where
a system's processes can be both alive and unresponsive, or otherwise
unhealthy. You should use liveness probes in order to notify Kubernetes
that your application's processes are unhealthy and should be restarted.
The Pod `template` for the `zk` StatefulSet specifies a liveness probe.
unhealthy. You should use liveness probes to notify Kubernetes
that your application's processes are unhealthy and it should restart them.
The Pod `template` for the `zk` `StatefulSet` specifies a liveness probe.
``
```yaml
livenessProbe:
@@ -805,11 +734,9 @@ The Pod `template` for the `zk` StatefulSet specifies a liveness probe.
timeoutSeconds: 5
```
The probe calls a simple bash script that uses the ZooKeeper `ruok` four letter
The probe calls a bash script that uses the ZooKeeper `ruok` four letter
word to test the server's health.
```bash
OK=$(echo ruok | nc 127.0.0.1 $1)
if [ "$OK" == "imok" ]; then
@@ -819,30 +746,25 @@ else
fi
```
In one terminal window, watch the Pods in the `zk` StatefulSet.
In one terminal window, use the following command to watch the Pods in the `zk` StatefulSet.
```shell
kubectl get pod -w -l app=zk
```
In another window, delete the `zkOk.sh` script from the file system of Pod `zk-0`.
In another window, using the following command to delete the `zkOk.sh` script from the file system of Pod `zk-0`.
```shell
kubectl exec zk-0 -- rm /usr/bin/zookeeper-ready
```
When the liveness probe for the ZooKeeper process fails, Kubernetes will
automatically restart the process for you, ensuring that unhealthy processes in
the ensemble are restarted.
```shell
kubectl get pod -w -l app=zk
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 0 1h
zk-1 1/1 Running 0 1h
@@ -853,25 +775,20 @@ zk-0 0/1 Running 1 1h
zk-0 1/1 Running 1 1h
```
### Testing for Readiness
Readiness is not the same as liveness. If a process is alive, it is scheduled
and healthy. If a process is ready, it is able to process input. Liveness is
a necessary, but not sufficient, condition for readiness. There are many cases,
a necessary, but not sufficient, condition for readiness. There are cases,
particularly during initialization and termination, when a process can be
alive but not ready.
If you specify a readiness probe, Kubernetes will ensure that your application's
processes will not receive network traffic until their readiness checks pass.
For a ZooKeeper server, liveness implies readiness. Therefore, the readiness
probe from the `zookeeper.yaml` manifest is identical to the liveness probe.
```yaml
readinessProbe:
exec:
@@ -883,25 +800,20 @@ probe from the `zookeeper.yaml` manifest is identical to the liveness probe.
timeoutSeconds: 5
```
Even though the liveness and readiness probes are identical, it is important
to specify both. This ensures that only healthy servers in the ZooKeeper
ensemble receive network traffic.
## Tolerating Node Failure
ZooKeeper needs a quorum of servers in order to successfully commit mutations
to data. For a three server ensemble, two servers must be healthy in order for
ZooKeeper needs a quorum of servers to successfully commit mutations
to data. For a three server ensemble, two servers must be healthy for
writes to succeed. In quorum based systems, members are deployed across failure
domains to ensure availability. In order to avoid an outage, due to the loss of an
domains to ensure availability. To avoid an outage, due to the loss of an
individual machine, best practices preclude co-locating multiple instances of the
application on the same machine.
By default, Kubernetes may co-locate Pods in a StatefulSet on the same node.
For the three server ensemble you created, if two servers reside on the same
node, and that node fails, the clients of your ZooKeeper service will experience
an outage until at least one of the Pods can be rescheduled.
By default, Kubernetes may co-locate Pods in a `StatefulSet` on the same node. For the three server ensemble you created, if two servers are on the same node, and that node fails, the clients of your ZooKeeper service will experience an outage until at least one of the Pods can be rescheduled.
You should always provision additional capacity to allow the processes of critical
systems to be rescheduled in the event of node failures. If you do so, then the
@@ -909,13 +821,13 @@ outage will only last until the Kubernetes scheduler reschedules one of the ZooK
servers. However, if you want your service to tolerate node failures with no downtime,
you should set `podAntiAffinity`.
Get the nodes for Pods in the `zk` Stateful Set.
Use the command below to get the nodes for Pods in the `zk` `StatefulSet`.
```shell{% raw %}
```shell
for i in 0 1 2; do kubectl get pod zk-$i --template {{.spec.nodeName}}; echo ""; done
``` {% endraw %}
```
All of the Pods in the `zk` StatefulSet are deployed on different nodes.
All of the Pods in the `zk` `StatefulSet` are deployed on different nodes.
```shell
kubernetes-minion-group-cxpk
@@ -923,7 +835,7 @@ kubernetes-minion-group-a5aq
kubernetes-minion-group-2g2d
```
This is because the Pods in the `zk` StatefulSet have a PodAntiAffinity specified.
This is because the Pods in the `zk` `StatefulSet` have a `PodAntiAffinity` specified.
```yaml
affinity:
@@ -954,7 +866,7 @@ The previous section showed you how to spread your Pods across nodes to survive
unplanned node failures, but you also need to plan for temporary node failures
that occur due to planned maintenance.
Get the nodes in your cluster.
Use this command to get the nodes in your cluster.
```shell
kubectl get nodes
@@ -963,55 +875,56 @@ kubectl get nodes
Use [`kubectl cordon`](/docs/user-guide/kubectl/{{page.version}}/#cordon) to
cordon all but four of the nodes in your cluster.
```shell{% raw %}
kubectl cordon < node name >
```{% endraw %}
```shell
kubectl cordon <node-name>
```
Get the `zk-pdb` PodDisruptionBudget.
Use this command to get the `zk-pdb` `PodDisruptionBudget`.
```shell
kubectl get pdb zk-pdb
```
The `max-unavailable` field indicates to Kubernetes that at most one Pod from
`zk` StatefulSet can be unavailable at any time.
`zk` `StatefulSet` can be unavailable at any time.
```shell
NAME MIN-AVAILABLE MAX-UNAVAILABLE ALLOWED-DISRUPTIONS AGE
zk-pdb N/A 1 1
```
In one terminal, watch the Pods in the `zk` StatefulSet.
In one terminal, use this command to watch the Pods in the `zk` `StatefulSet`.
```shell
kubectl get pods -w -l app=zk
```
In another terminal, get the nodes that the Pods are currently scheduled on.
In another terminal, use this command to get the nodes that the Pods are currently scheduled on.
```shell{% raw %}
```shell
for i in 0 1 2; do kubectl get pod zk-$i --template {{.spec.nodeName}}; echo ""; done
kubernetes-minion-group-pb41
kubernetes-minion-group-ixsl
kubernetes-minion-group-i4c4
{% endraw %}
```
Use [`kubectl drain`](/docs/user-guide/kubectl/{{page.version}}/#drain) to cordon and
drain the node on which the `zk-0` Pod is scheduled.
```shell {% raw %}
```shell
kubectl drain $(kubectl get pod zk-0 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-local-data
node "kubernetes-minion-group-pb41" cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-minion-group-pb41, kube-proxy-kubernetes-minion-group-pb41; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-o5elz
pod "zk-0" deleted
node "kubernetes-minion-group-pb41" drained
{% endraw %}```
```
As there are four nodes in your cluster, `kubectl drain`, succeeds and the
`zk-0` is rescheduled to another node.
```
```shell
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 2 1h
zk-1 1/1 Running 0 1h
@@ -1028,22 +941,22 @@ zk-0 0/1 Running 0 51s
zk-0 1/1 Running 0 1m
```
Keep watching the StatefulSet's Pods in the first terminal and drain the node on which
Keep watching the `StatefulSet`'s Pods in the first terminal and drain the node on which
`zk-1` is scheduled.
```shell{% raw %}
```shell
kubectl drain $(kubectl get pod zk-1 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-local-data "kubernetes-minion-group-ixsl" cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-minion-group-ixsl, kube-proxy-kubernetes-minion-group-ixsl; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-voc74
pod "zk-1" deleted
node "kubernetes-minion-group-ixsl" drained
{% endraw %}```
```
The `zk-1` Pod can not be scheduled. As the `zk` StatefulSet contains a
PodAntiAffinity rule preventing co-location of the Pods, and as only
two nodes are schedulable, the Pod will remain in a Pending state.
The `zk-1` Pod cannot be scheduled because the `zk` `StatefulSet` contains a `PodAntiAffinity` rule preventing co-location of the Pods, and as only two nodes are schedulable, the Pod will remain in a Pending state.
```shell
kubectl get pods -w -l app=zk
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 2 1h
zk-1 1/1 Running 0 1h
@@ -1069,19 +982,19 @@ zk-1 0/1 Pending 0 0s
Continue to watch the Pods of the stateful set, and drain the node on which
`zk-2` is scheduled.
```shell{% raw %}
```shell
kubectl drain $(kubectl get pod zk-2 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-local-data
node "kubernetes-minion-group-i4c4" cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-minion-group-i4c4, kube-proxy-kubernetes-minion-group-i4c4; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog
WARNING: Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog; Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-minion-group-i4c4, kube-proxy-kubernetes-minion-group-i4c4
There are pending pods when an error occurred: Cannot evict pod as it would violate the pod's disruption budget.
pod/zk-2
{% endraw %}```
```
Use `CTRL-C` to terminate to kubectl.
You can not drain the third node because evicting `zk-2` would violate `zk-budget`. However,
the node will remain cordoned.
You cannot drain the third node because evicting `zk-2` would violate `zk-budget`. However, the node will remain cordoned.
Use `zkCli.sh` to retrieve the value you entered during the sanity test from `zk-0`.
@@ -1089,9 +1002,9 @@ Use `zkCli.sh` to retrieve the value you entered during the sanity test from `zk
kubectl exec zk-0 zkCli.sh get /hello
```
The service is still available because its PodDisruptionBudget is respected.
The service is still available because its `PodDisruptionBudget` is respected.
```
```shell
WatchedEvent state:SyncConnected type:None path:null
world
cZxid = 0x200000002
@@ -1111,6 +1024,7 @@ Use [`kubectl uncordon`](/docs/user-guide/kubectl/{{page.version}}/#uncordon) to
```shell
kubectl uncordon kubernetes-minion-group-pb41
node "kubernetes-minion-group-pb41" uncordoned
```
@@ -1118,6 +1032,7 @@ node "kubernetes-minion-group-pb41" uncordoned
```shell
kubectl get pods -w -l app=zk
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 2 1h
zk-1 1/1 Running 0 1h
@@ -1146,14 +1061,15 @@ zk-1 1/1 Running 0 13m
Attempt to drain the node on which `zk-2` is scheduled.
```shell{% raw %}
```shell
kubectl drain $(kubectl get pod zk-2 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-local-data
node "kubernetes-minion-group-i4c4" already cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-minion-group-i4c4, kube-proxy-kubernetes-minion-group-i4c4; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog
pod "heapster-v1.2.0-2604621511-wht1r" deleted
pod "zk-2" deleted
node "kubernetes-minion-group-i4c4" drained
{% endraw %}```
```
This time `kubectl drain` succeeds.
@@ -1161,22 +1077,20 @@ Uncordon the second node to allow `zk-2` to be rescheduled.
```shell
kubectl uncordon kubernetes-minion-group-ixsl
node "kubernetes-minion-group-ixsl" uncordoned
```
You can use `kubectl drain` in conjunction with PodDisruptionBudgets to ensure that your service
remains available during maintenance. If drain is used to cordon nodes and evict pods prior to
taking the node offline for maintenance, services that express a disruption budget will have that
budget respected. You should always allocate additional capacity for critical services so that
their Pods can be immediately rescheduled.
You can use `kubectl drain` in conjunction with `PodDisruptionBudgets` to ensure that your services remain available during maintenance. If drain is used to cordon nodes and evict pods prior to taking the node offline for maintenance, services that express a disruption budget will have that budget respected. You should always allocate additional capacity for critical services so that their Pods can be immediately rescheduled.
{% endcapture %}
{% capture cleanup %}
* Use `kubectl uncordon` to uncordon all the nodes in your cluster.
* You will need to delete the persistent storage media for the PersistentVolumes
used in this tutorial. Follow the necessary steps, based on your environment,
storage configuration, and provisioning method, to ensure that all storage is
reclaimed.
{% endcapture %}
{% include templates/tutorial.md %}
- Use `kubectl uncordon` to uncordon all the nodes in your cluster.
- You will need to delete the persistent storage media for the PersistentVolumes
used in this tutorial. Follow the necessary steps, based on your environment,
storage configuration, and provisioning method, to ensure that all storage is
reclaimed.
{% endcapture %}
{% include templates/tutorial.md %}