Apache Karaf Cellar 3.x
Architecture Guide
1. Architecture Overview
The core concept behind Karaf Cellar is that each node can be a part of one or more groups that provide the node distributed memory for keeping data (e.g. configuration, features information, other) and a topic which is used to exchange events with the rest of the group nodes.
Each group comes with a configuration, which defines which events are to be broadcasted and which are not. Whenever a local change occurs to a node, the node will read the setup information of all the groups that it belongs to and broadcasts the event to the groups that are whitelisted to the specific event.
The broadcast operation happens via a distributed topic provided by the group. For the groups that the broadcast reaches, the distributed configuration data will be updated so that nodes that join in the future can pickup the change.
1. Introduction
1.1. Karaf Cellar use cases
The first purpose of Cellar is to synchronize the state of several Karaf instances (named nodes).
Cellar provides dedicated shell commands and MBeans to administrate the cluster, and manipulate the resources on the cluster.
It’s also possible to enable local resources listeners: these listeners broadcast local resource changes as cluster events. Please note that this behavior is disabled by default as it can have side effects (especially when a node is stopped). Enabling listeners is at your own risk.
The nodes list could be discovered (using uni-cast or multicast), or "statically" defined (using a couple host name or IP and port list).
Cellar is able to synchronize:
- bundles (remote, local, or from an OBR)
- config
- features
- event-admin
Optionally, Cellar also support synchronization of OSGi Event Admin, OBR (URLs and bundles).
The second purpose is to provide a Distributed OSGi run time. It means that using Cellar, you are able to call an OSGi service located on a remote instance. See the [Transport and DOSGi] section of the user guide.
Finally, Cellar also provides "run time clustering" by providing dedicated feature like:
- HTTP load balancing
- HTTP sessions replication
- log centralization
Please, see the sections dedicated to those features.
1.2. Cross topology
This is the default Cellar topology. Cellar is installed on every nodes, each node has the same function.
It means that you can perform actions on any node, it will be broadcast to all others nodes.
1.3. Star topology
In this topology, if Cellar is installed on all nodes, you perform actions only on one specific node (the "manager").
To do that, the "manager" is a standard Cellar node, and the event producing is disable on all others nodes (
cluster:producer-stop on all "managed" nodes).
Like this, only the "manager" will send event to the nodes (which are able to consumer and handle), but no event can be produced on the nodes.
2. Installation
This chapter describes how to install Apache Karaf Cellar into your existing Karaf based installation.
2.1. Pre-Installation Requirements
Cellar is installed on running Karaf instances.
Cellar is provided as a Karaf features descriptor. The easiest way to install is just to have an internet connection from the Karaf running instance.
3. Deploy Cellar
This chapter describes how to deploy and start Cellar into a running Apache Karaf instance. This chapter assumes that you already know Apache Karaf basics, especially the notion of features and shell usage.
3.1. Registering Cellar features
Karaf Cellar is provided as a Karaf features XML descriptor.
Simply register the Cellar feature URL in your Karaf instance:
karaf@root()> feature:repo-add mvn:org.apache.karaf.cellar/apache-karaf-cellar/3.0.3/xml/features
Adding feature url mvn:org.apache.karaf.cellar/apache-karaf-cellar/3.0.3/xml/features
Now you have Cellar features available in your Karaf instance:
karaf@root()> feature:list |grep -i cellar
cellar-core | 3.0.3 | | karaf-cellar-3.0.3 | Karaf clustering core
hazelcast | 3.4.2 | | karaf-cellar-3.0.3 | In memory data grid
cellar-hazelcast | 3.0.3 | | karaf-cellar-3.0.3 | Cellar implementation based on Hazelcast
cellar-config | 3.0.3 | | karaf-cellar-3.0.3 | ConfigAdmin cluster support
cellar-features | 3.0.3 | | karaf-cellar-3.0.3 | Karaf features cluster support
cellar-bundle | 3.0.3 | | karaf-cellar-3.0.3 | Bundle cluster support
cellar-shell | 3.0.3 | | karaf-cellar-3.0.3 | Cellar shell support
cellar | 3.0.3 | | karaf-cellar-3.0.3 | Karaf clustering
cellar-dosgi | 3.0.3 | | karaf-cellar-3.0.3 | DOSGi support
cellar-obr | 3.0.3 | | karaf-cellar-3.0.3 | OBR cluster support
cellar-eventadmin | 3.0.3 | | karaf-cellar-3.0.3 | OSGi events broadcasting in clusters
cellar-cloud | 3.0.3 | | karaf-cellar-3.0.3 | Cloud blobstore support in clusters
cellar-webconsole | 3.0.3 | | karaf-cellar-3.0.3 | Cellar plugin for Karaf WebConsole
3.2. Starting Cellar
To start Cellar in your Karaf instance, you only need to install the Cellar feature:
karaf@root()> feature:install cellar
You can now see the Cellar components (bundles) installed:
karaf@root()> la|grep -i cellar
80 | Active | 30 | 3.0.3 | Apache Karaf :: Cellar :: Core
81 | Active | 31 | 3.0.3 | Apache Karaf :: Cellar :: Utils
82 | Active | 33 | 3.0.3 | Apache Karaf :: Cellar :: Hazelcast
83 | Active | 40 | 3.0.3 | Apache Karaf :: Cellar :: Shell
84 | Active | 40 | 3.0.3 | Apache Karaf :: Cellar :: Config
85 | Active | 40 | 3.0.3 | Apache Karaf :: Cellar :: Bundle
86 | Active | 40 | 3.0.3 | Apache Karaf :: Cellar :: Features
And Cellar cluster commands are now available:
karaf@root()> cluster:<TAB>
3.3. Optional features
Optionally, you can install additional features.
The cellar-event feature adds support of OSGi Event Admin on the cluster:
karaf@root()> feature:install cellar-event
The cellar-obr feature adds support of OBR sync on the cluster:
karaf@root()> feature:install cellar-obr
The cellar-dosgi feature adds support of DOSGi (Distributed OSGi):
karaf@root()> feature:install cellar-dosgi
The cellar-cloud feature adds support of cloud blob-store, allowing to use instances located on a cloud provider:
karaf@root()> feature:install cellar-cloud
4. Core runtime and Hazelcast
Cellar uses Hazelcast as cluster engine.
When you install the cellar feature, a hazelcast feature is automatically installed, providing the
etc/hazelcast.xmlconfiguration file.
The
etc/hazelcast.xml configuration file contains all the core configuration, especially: * the Hazelcast cluster identifiers (group name and password) * network discovery and security configuration4.1. Hazelcast cluster identification
The
<group/> element in the etc/hazelcast.xml defines the identification of the Hazelcast cluster: <group>
<name>cellar</name>
<password>pass</password>
</group>
All Cellar nodes have to use the same name and password (to be part of the same Hazelcast cluster).
4.2. Network
The
<network/> element in the etc/hazelcast.xml contains all the network configuration.
First, it defines the port numbers used by Hazelcast:
<port auto-increment="true" port-count="100">5701</port>
<outbound-ports>
<!--
Allowed port range when connecting to other nodes.
0 or * means use system provided port.
-->
<ports>0</ports>
</outbound-ports>
Second, it defines the mechanism used to discover the Cellar nodes: it’s the
<join/> element.
By default, Hazelcast uses unicast.
You can also use multicast (enabled by default in Cellar):
<multicast enabled="true">
<multicast-group>224.2.2.3</multicast-group>
<multicast-port>54327</multicast-port>
</multicast>
<tcp-ip enabled="false"/>
<aws enabled="false"/>
Instead of using multicast, you can also explicitly define the host names (or IP addresses) of the different Cellar nodes:
<multicast enabled="false"/>
<tcp-ip enabled="true"/>
<aws enabled="false"/>
By default, it will bind to all interfaces on the node machine. It’s possible to specify a interface:
<multicast enabled="false"/>
<tcp-ip enabled="true">
<interface>127.0.0.1</interface>
</tcp-ip>
<aws enabled="false"/>
NB: in previous Hazelcast versions (especially the one used by Cellar 2.3.x), it was possible to have multicast and tcp-ip enabled in the same time. In Hazelcast 3.3.x (the version currently used by Cellar 3.0.x), only one discover mechanism can be enabled at a time. Cellar uses multicast by default (tcp-ip is disabled). If your network or network interface don’t support multicast, you have to enable tcp-ip and disable multicast.
You can also discover nodes located on a Amazon instance:
<multicast enabled="false"/>
<tcp-ip enabled="false"/>
<aws enabled="true">
<access-key>my-access-key</access-key>
<secret-key>my-secret-key</secret-key>
<!--optional, default is us-east-1 -->
<region>us-west-1</region>
<!--optional, default is ec2.amazonaws.com. If set, region shouldn't be set as it will override this property -->
<host-header>ec2.amazonaws.com</host-header>
<!-- optional, only instances belonging to this group will be discovered, default will try all running instances -->
<security-group-name>hazelcast-sg</security-group-name>
<tag-key>type</tag-key>
<tag-value>hz-nodes</tag-value>
</aws>
Third, you can specific on which network interface the cluster is running (whatever the discovery mechanism used). By default, Hazelcast listens on all interfaces (0.0.0.0). But you can specify an interface:
<interfaces enabled="true">
<interface>10.10.1.*</interface>
</interfaces>
Finally, you can also enable security transport on the cluster. Two modes are supported: * SSL:
<ssl enabled="true"/>
- Symmetric Encryption:
<symmetric-encryption enabled="true">
<!--
encryption algorithm such as
DES/ECB/PKCS5Padding,
PBEWithMD5AndDES,
AES/CBC/PKCS5Padding,
Blowfish,
DESede
-->
<algorithm>PBEWithMD5AndDES</algorithm>
<!-- salt value to use when generating the secret key -->
<salt>thesalt</salt>
<!-- pass phrase to use when generating the secret key -->
<password>thepass</password>
<!-- iteration count to use when generating the secret key -->
<iteration-count>19</iteration-count>
</symmetric-encryption>
Cellar provides additional discovery mechanisms, See Discovery Service (jclouds and kubernetes) section for details.
5. Cellar nodes
This chapter describes the Cellar nodes manipulation commands.
5.1. Nodes identification
When you installed the Cellar feature, your Karaf instance became automatically a Cellar cluster node, and hence tries to discover the others Cellar nodes.
You can list the known Cellar nodes using the list-nodes command:
karaf@root()> cluster:node-list
| Id | Host Name | Port
-------------------------------------
x | node2:5702 | node2 | 5702
| node1:5701 | node1 | 5701
The starting x indicates that it’s the Karaf instance on which you are logged on (the local node).
NB: if you don’t see the other nodes there (whereas they should be there), it’s probably due to a network issue. By default, Cellar uses multicast to discover the nodes. If your network or network interface don’t support multicast, you have to switch to tcp-ip instead of multicast. See [Core Configuration|hazelcast] for details.
NB: in Cellar 2.3.x, Cellar used both multicast and tcp-ip by default. Due to a change in Hazelcast, it’s no more possible to have both. Now, in Cellar 3.0.x, the default configuration is multicast enabled, tcp-ip disabled. See [Core Configuration|hazelcast] for details.
5.2. Testing nodes
You can ping a node to test it:
karaf@root()> cluster:node-ping node1:5701
PING node1:5701
from 1: req=node1:5701 time=11 ms
from 2: req=node1:5701 time=12 ms
from 3: req=node1:5701 time=13 ms
from 4: req=node1:5701 time=7 ms
from 5: req=node1:5701 time=12 ms
5.3. Node Components: listener, producer, handler, consume, and synchronize
A Cellar node is actually a set of components, each component is dedicated to a special purpose.
The
etc/org.apache.karaf.cellar.node.cfg configuration file is dedicated to the configuration of the local node. It’s where you can control the status of the different components.5.4. Synchronize and sync policy
A synchronize is invoked when you:
- Cellar starts
- a node joins a cluster group (see [groups] for details about cluster groups)
- you explicitly call the cluster:sync command
We have a synchronize per resource: feature, bundle, config, obr (optional).
Cellar supports three sync policies:
- cluster (default): if the node is the first one in the cluster, it pushes its local state to the cluster, else if it’s not the first node in the cluster, the node will update its local state with the cluster one (meaning that the cluster is the master).
- node: in this case, the node is the master, it means that the cluster state will be overwritten by the node state.
- disabled: in this case, it means that the synchronize is not used at all, meaning the node or the cluster are not updated at all (at sync time).
You can configure the sync policy (for each resource, and each cluster group) in the etc/org.apache.karaf.cellar.groups.cfg configuration file:
default.bundle.sync = cluster
default.config.sync = cluster
default.feature.sync = cluster
default.obr.urls.sync = cluster
The
cluster:sync command allows you to "force" the sync:karaf@node1()> cluster:sync
Synchronizing cluster group default
bundle: done
config: done
feature: done
obr.urls: No synchronizer found for obr.urls
It’s also possible to sync only a resource using:
- -b (--bundle) for bundle
- -f (--feature) for feature
- -c (--config) for configuration
- -o (--obr) for OBR URLs
or a given cluster group using the -g (--group) option.
6. Cellar groups
You can define groups in Cellar. A group allows you to define specific nodes and resources that are to be working together. This permits some nodes (those outside the group) not to need to sync’ed with changes of a node within a group.
By default, the Cellar nodes go into the default group:
karaf@root()> cluster:group-list
| Group | Members
-----------------------------------------------
x | default | node2:5702 node1:5701(x)
The x indicates a local group. A local group is a group containing the local node (where we are connected).
6.1. New group
You can create a new group using the group-create command:
karaf@root()> cluster:group-create test
For now, the test group hasn’t any nodes:
karaf@node1()> cluster:group-list
| Group | Members
-----------------------------------------------
x | default | node2:5702 node1:5701(x)
| test |
7. DOSGi and Transport
DOSGi (Distributed OSGi) enables the distribution of OSGi services across the Cellar nodes.
The purpose of the Cellar DOSGi is to leverage the Cellar resources (Hazelcast instances, distributed map, etc), and to be very easy to use.
DOSGi is provided by installing the optional feature cellar-dosgi.
To be available and visible for the others nodes, the OSGi service should only have the
service.exported.interfacesproperty:<service ref="MyService" interface="my.interface">
<service-properties>
<entry key="service.exported.interfaces" value="*"/>
</service-properties>
</service>
You can see all OSGi services "flagged" as distributed (available for the nodes) using the
cluster:list-servicecommand:karaf@root()> cluster:service-list
A "client" bundle could use this service. If the service is not available locally, Cellar will "route" the service call to the remote remote containing the service.
8. Discovery Services
The Discovery Services allow you to use third party libraries to discover the nodes member of the Cellar cluster.
8.1. jClouds
Cellar relies on Hazelcast (http://www.hazelcast.com) in order to discover cluster nodes. This can happen either by using unicast, multicast or specifying the ip address of each node. See the Core Configuration section for details.
Unfortunately multicast is not allowed in most IaaS providers and the alternative of specifying all IP addresses creates maintenance difficulties, especially since in most cases the addresses are not known in advance.
Cellar solves this problem using a cloud discovery service powered by jclouds (http://jclouds.apache.org).
8.1.1. Cloud discovery service
Most cloud providers provide cloud storage among other services. Cellar uses the cloud storage via jclouds, in order to determine the IP addresses of each node so that Hazelcast can find them.
This approach is also called blackboard and refers to the process where each node registers itself in a common storage are so that other nodes know its existence.
8.1.2. Installing Cellar cloud discovery service
To install the cloud discovery service simply install the appropriate jclouds provider and then install cellar-cloud feature. Amazon S3 is being used here for this example, but the below applies to any provider supported by jclouds.
karaf@root()> feature:install jclouds-aws-s3
karaf@root()> feature:install cellar-cloud
Once the feature is installed, you’re required to create a configuration that contains credentials and the type of the cloud storage (aka blobstore). To do that add a configuration file under the etc folder with the name
org.apache.karaf.cellar.cloud-<provider>.cfg and place the following information there:provider=aws-s3 (this varies according to the blobstore provider)
identity=<the identity of the blobstore account>
credential=<the credential/password of the blobstore account)
container=<the name of the bucket>
validity=<the amount of time an entry is considered valid, after that time the entry is removed>
For instance, you can create
etc/org.apache.karaf.cellar.cloud-mycloud.cfg containing:provider=aws-s3
identity=username
credential=password
container=cellar
validity=360000
NB: you can find the cloud providers supported by jclouds herehttp://repo1.maven.org/maven2/org/apache/jclouds/provider/. You have to install the corresponding jclouds feature for the provider.
After creating the file the service will check for new nodes. If new nodes are found the Hazelcast instance configuration will be updated and the instance restarted.
8.1.3. Kubernetes & docker.io
Kubernetes is an open source orchestration system for docker.io containers. It handles scheduling onto nodes in a compute cluster and actively manages workloads to ensure that their state matches the users declared intentions. Using the concepts of "labels", "pods", "replicationControllers" and "services", it groups the containers which make up an application into logical units for easy management and discovery. Following the aforementioned concept will most likely change how you package and provision your Karaf based applications. For instance, you will eventually have to provide a Docker image with a pre-configured Karaf, KAR files in deployment folder, etc. so that your Kubernetes container may bootstrap everything on boot.
The Cellar Kubernetes discovery service is a great complement to the Karaf docker.io feature (allowing you to easily create and manage docker.io images in and for Karaf).
8.1.4. Kubernetes discovery service
In order to determine the IP address of each node, so that Hazelcast can connect to them, the Kubernetes discovery service queries the Kubernetes API for containers labeled with the
pod.label.key and pod.label.key specified inetc/org.apache.karaf.cellar.kubernetes-name.cfg. The name in etc/org.apache.karaf.cellar.kubernetes-name.cfg is a name of the choice. It allows you to create multiple Kubernetes discovery services. Thanks to that, the Cellar nodes can be discovered on different Kubernetes.
So, you must be sure to label your containers (pods) accordingly.
After a Cellar node starts up, Kubernetes discovery service will configure Hazelcast with currently running Cellar nodes. Since Hazelcast follows a peer-to-peer all-shared topology, whenever nodes come up and down, the cluster will remain up-to-date.
11.1.5. Installing Kubernetes discovery service
To install the Kubernetes discovery service, simply install cellar-kubernetes feature.
karaf@root()> feature:install cellar-kubernetes
Once the cellar-kubernetes feature is installed, you have to create the Kubernetes provider configuration file. If you have multiple Kubernetes instances, you create one configuration file per instance.
For instance, you can create
etc/org.apache.karaf.cellar.kubernetes-myfirstcluster.cfg containing:host=localhost
port=8080
pod.label.key=name
pod.label.value=cellar
and another one
etc/org.apache.karaf.cellar.kubernetes-mysecondcluster.cfg containing:host=192.168.134.2
port=8080
pod.label.key=name
pod.label.value=cellar
In case you change the file, the discovery service will check again for new nodes. If new nodes are found, Hazelcast configuration will be updated and the instance restarted.
Reference
http://karaf.apache.org/
No comments:
Post a Comment