Deploy Alluxio on a Cluster with HA

Slack Docker Pulls

Overview

An Alluxio cluster with High Availability (HA) is achieved by running multiple Alluxio master processes on different nodes in the system. One master is elected as the leading master which serves all workers and clients as the primary point of contact. The other masters act as standby masters and maintain the same file system state as the leading master by reading a shared journal. Standby masters do not serve any client or worker requests; however, if the leading master fails, one standby master will automatically be elected as the new leading master. Once the new leading master starts serving, Alluxio clients and workers proceed as usual. During the failover to a standby master, clients may experience brief delays or transient errors.

The major challenges to achieving high-availability are maintaining a shared file system state across service restarts and maintaining consensus among masters about the identity of the leading master after failover. In Alluxio 2.0, there are two different ways to achieve these two goals:

  1. Raft-based Journal: Uses an internal replicated state machine based on the Raft protocol to both store the file system journal and run leader elections. This approach is introduced in Alluxio 2.0 and requires no dependency on external services.
  2. Zookeeper with a shared Journal: Uses an external Zookeeper service for leader elections in conjunction with a shared storage (e.g. HDFS) for the shared journal. See journal management documentation for more information about choosing and configuring Alluxio journal system.

Prerequisites

  • To deploy an Alluxio cluster, first download the pre-compiled Alluxio binary file, extract the tarball and copy the extracted directory to all nodes (including nodes running masters and workers). Alluxio Enterprise may also be launched through the AWS marketplace via an AMI which contains the extracted tarball.
  • Enable SSH login without password from all master nodes to all worker nodes. You can add a public SSH key for the host into ~/.ssh/authorized_keys. See this tutorial for more details.
  • TCP traffic across all nodes is allowed. For basic functionality make sure RPC port (default :19998) is open on all nodes.

Basic Setup

Raft-based Embedded Journal

The minimal configuration to set up a HA cluster is to give the embedded journal addresses to all nodes inside the cluster. On each Alluxio node, create the conf/alluxio-site.properties configuration file from the template.

$ cp conf/alluxio-site.properties.template conf/alluxio-site.properties

Add the following properties to the conf/alluxio-site.properties file:

alluxio.master.hostname=<MASTER_HOSTNAME> # Only needed on master node
<!-- ALLUXIO CS ADD -->
alluxio.master.journal.type=EMBEDDED
<!-- ALLUXIO CS END -->
alluxio.master.embedded.journal.addresses=<EMBEDDED_JOURNAL_ADDRESS>

Explanation:

  • The first property alluxio.master.hostname=<MASTER_HOSTNAME> is required on each master node to be its own externally visible hostname. This is required on each individual component of the master quorum to have its own address set. On worker nodes, this parameter will be ignored. Examples include alluxio.master.hostname=1.2.3.4, alluxio.master.hostname=node1.a.com.
  • Property alluxio.master.journal.type=EMBEDDED uses a journal embedded in the masters. Note that Zookeeper cannot work with journal type EMBEDDED.
  • The second property alluxio.master.embedded.journal.addresses sets the sets of masters to participate Alluxio’s internal leader election and determine the leading master. The default embedded journal port is 19200. An example: alluxio.master.embedded.journal.addresses=master_hostname_1:19200,master_hostname_2:19200,master_hostname_3:19200

Note that embedded journal feature relies on Ratis which uses leader election based on the Raft protocol and has its own format for storing journal entries. The built-in leader election cannot work with Zookeeper since the journal formats between these configurations may not match. Enabling embedded journal enables Alluxio’s internal leader election. See embedded journal configuration documentation for more details and alternative ways to set up HA cluster with internal leader election.

Zookeeper and Shared Journal Storage

When using Zookeeper for HA there are additional prerequisites:

  1. A ZooKeeper cluster. Alluxio masters use ZooKeeper for leader election, and Alluxio clients and workers use ZooKeeper to inquire about the identity of the current leading master.
  2. A shared storage system on which to place the journal (accessible by all Alluxio masters). The leading master writes to the journal on this shared storage system, while the standby masters continually replay the journal entries to stay up-to-date. The journal storage system is recommended to be:
    • Highly available. All metadata modifications on the master requires writing to the journal, so any downtime of the journal storage system will directly impact the Alluxio master availability.
    • A filesystem, not an object store. The Alluxio master writes to journal files to this storage system, and utilizes filesystem operations such as rename and flush. Object stores do not support these operations, and/or perform them slowly, so when the journal is stored on an object store, the Alluxio master operation throughput is significantly reduced.

The minimal configuration parameters which must be set are:

alluxio.zookeeper.enabled=true
alluxio.zookeeper.address=<ZOOKEEPER_ADDRESS>
<!-- ALLUXIO CS REMOVE -->
<!-- alluxio.master.journal.type=UFS -->
<!-- ALLUXIO CS END -->
alluxio.master.journal.folder=<JOURNAL_URI>

Explanation:

  • alluxio.zookeeper.enabled=true enables the HA mode for the masters, and informs workers that HA mode is enabled.
  • alluxio.zookeeper.address=<ZOOKEEPER_ADDRESS> sets the ZooKeeper address when alluxio.zookeeper.enabled is enabled. The HA masters use ZooKeeper for leader election, and the workers use ZooKeeper to discover the leading master. Multiple ZooKeeper addresses can be specified by delimiting with commas. Examples include alluxio.zookeeper.address=1.2.3.4:2181, alluxio.zookeeper.address=zk1:2181, zk2:2181,zk3:2181
  • alluxio.master.journal.folder=<JOURNAL_URI> sets the URI of the shared journal location for the Alluxio leading master to write the journal to, and for standby masters to replay journal entries from. This shared shared storage system must be accessible by all master nodes. Examples include alluxio.master.journal.folder=hdfs://1.2.3.4:9000/alluxio/journal/ or alluxio.master.journal.folder=/mnt/nfs/journal/.

Make sure all master nodes and all worker nodes have configured their respective conf/alluxio-site.properties configuration file appropriately.

Once all the Alluxio masters and workers are configured in this way, Alluxio is ready to be formatted and started.

Advanced Zookeeper setup

For clusters with large namespaces, increased CPU overhead on leader could cause delays on Zookeeper client heartbeats. For this reason, we recommend setting Zookeeper client session timeout to at least 2 minutes on large clusters with namespace size more than several hundred million of files.

  • alluxio.zookeeper.session.timeout=120s
    • The Zookeeper server’s min/max session timeout values must also be configured as such to allow this timeout. The defaults requires that the timeout be a minimum of 2 times the tickTime (as set in the server configuration) and a maximum of 20 times the tickTime. You could also manually configure minSessionTimeout and maxSessionTimeout.

Alluxio supports pluggable error handling policy on Zookeeper leader election.

  • alluxio.zookeeper.leader.connection.error.policy specifies how connection errors are handled. It can be either SESSION or STANDARD. It is set SESSION as default.

  • The SESSION policy makes use of Zookeeper sessions to determine whether leader state is dirty. This means suspended connections won’t trigger stepping down of a current leader as long as it was able to reestablish the zookeeper connection with the same session. It provides more stability in maintaining the leadership state.
  • The STANDARD policy treats any interruption to zookeeper server as an error. Thus leader will step down upon missing a heartbeat, even though its internal zookeeper session was still intact with the zookeeper server. It provides more security against bugs and issues in zookeeper setup.

Start an Alluxio Cluster with HA

Format Alluxio

Before Alluxio can be started for the first time, the Alluxio master journal and worker storage must be formatted.

Formatting the journal will delete all metadata from Alluxio. Formatting the worker storage will delete all data from the configured Alluxio storage. However, the data in under storage will be untouched.

On all the Alluxio master nodes, list all the worker hostnames in the conf/workers file, and list all the masters in the conf/masters file. This will allow alluxio scripts to run operations on the cluster nodes. format Alluxio cluster with the following command in one of the master nodes:

$ ./bin/alluxio format

Launch Alluxio

In one of the master nodes, start the Alluxio cluster with the following command:

$ ./bin/alluxio-start.sh all SudoMount

This will start Alluxio masters on all the nodes specified in conf/masters, and start the workers on all the nodes specified in conf/workers. Argument SudoMount indicates to mount the RamFS on each worker using sudo privilege, if it is not already mounted. On MacOS, make sure your terminal has full disk access (tutorial here).

Verify Alluxio Cluster

To verify that Alluxio is running, you can visit the web UI of the leading master. To determine the leading master, run:

$ ./bin/alluxio fs masterInfo

Then, visit http://<LEADER_HOSTNAME>:19999 to see the status page of the Alluxio leading master.

Alluxio comes with a simple program that writes and reads sample files in Alluxio. Run the sample program with:

$ ./bin/alluxio runTests

Access an Alluxio Cluster with HA

When an application interacts with Alluxio in HA mode, the client must know about the connection information of Alluxio HA cluster, so that the client knows how to discover the Alluxio leading master. The following sections list three ways to specify the HA Alluxio service address on the client side.

Specify Alluxio Service in Configuration Parameters or Java Options

Users can pre-configure the service address of an Alluxio HA cluster in environment variables or site properties, and then connect to the service using an Alluxio URI such as alluxio:///path. For example, with Alluxio connection information in core-site.xml of Hadoop, Hadoop CLI can connect to the Alluxio cluster.

$ hadoop fs -ls alluxio:///directory

Depending on the different approaches to achieve HA, different properties are required:

If using embedded journal, set alluxio.master.rpc.addresses.

alluxio.master.rpc.addresses=master_hostname_1:19998,master_hostname_2:19998,master_hostname_3:19998

Or specify the properties in Java option. For example, for Spark applications, add the following to spark.executor.extraJavaOptions and spark.driver.extraJavaOptions:

-Dalluxio.master.rpc.addresses=master_hostname_1:19998,master_hostname_2:19998,master_hostname_3:19998

If using Zookeeper, set the following Zookeeper related properties

alluxio.zookeeper.enabled=true
alluxio.zookeeper.address=<ZOOKEEPER_ADDRESS>

Note that, the ZooKeeper address (alluxio.zookeeper.address) must be specified when alluxio.zookeeper.enabled is enabled and vise versa. Multiple ZooKeeper addresses can be specified by delimiting with commas.

Specify Alluxio Service with URL Authority

Users can also fully specify the HA cluster information in the URI to connect to an Alluxio HA cluster. Configuration derived from the HA authority takes precedence over all other forms of configuration, e.g. site properties or environment variables.

  • When using embedded journal, use alluxio://master_hostname_1:19998, master_hostname_2:19998,master_hostname_3:19998/path
  • When using Zookeeper leader election, use alluxio://zk@<ZOOKEEPER_ADDRESS>/path.

For many applications (e.g., Hadoop, Hive and Flink), you can use a comma as the delimiter for multiple addresses in the URI, like alluxio://master_hostname_1:19998,master_hostname_2:19998,master_hostname_3:19998/path and alluxio://zk@zkHost1:2181,zkHost2:2181,zkHost3:2181/path.

For some other applications (e.g., Spark) where comma is not accepted inside a URL authority, you need to use semicolons as the delimiter for multiple addresses, like alluxio://master_hostname_1:19998;master_hostname_2:19998;master_hostname_3:19998 and alluxio://zk@zkHost1:2181;zkHost2:2181;zkHost3:2181/path.

Specify Alluxio Service with logical URL Authority

Some frameworks may not accept either of these ways to connect to a highly available Alluxio HA cluster, so Alluxio also supports connecting to an Alluxio HA cluster via a logical name. In order to use logical names, the following configuration options need to be set in your environment variables or site properties.

Use logical name when using embedded journal

If you are using embedded journal, you need to configure the following configuration options and connect to the highly available alluxio node via alluxio://ebj@[logical-name] , for example alluxio://ebj@my-alluxio-cluster.

  • alluxio.master.nameservices.[logical-name] unique identifier for each alluxio master node

A comma-separated ID of the alluxio master node that determine all the alluxio master nodes in the cluster. For example, if you previously used my-alluxio-cluster as the logical name and wanted to use master1,master2,master3 as individual IDs for each alluxio master, you configure this as such:

alluxio.master.nameservices.my-alluxio-cluster=master1,master2,master3
  • alluxio.master.rpc.address.[logical name]. [master node ID] RPC Address for each alluxio master node

For each alluxio master node previously configured, set the full address of each alluxio master node, for example:

alluxio.master.rpc.address.my-alluxio-cluster.master1=master1:19998
alluxio.master.rpc.address.my-alluxio-cluster.master2=master2:19998
alluxio.master.rpc.address.my-alluxio-cluster.master3=master3:19998

Use logical name when using Zookeeper

If you are using zookeeper for leader election, you need to configure the following values and connect to the highly available alluxio node via alluxio://zk@[logical-name] , for example alluxio://zk@my-alluxio-cluster.

  • alluxio.master.zookeeper.nameservices.[logical-name] unique identifier for each Zookeeper node

A comma-separated zookeeper node ID that determine all the Zookeeper nodes in the cluster. For example, if you previously used my-alluxio-cluster as the logical name and wanted to use node1,node2,node3 as individual IDs for each Zookeeper, you would configure this as such:

alluxio.master.zookeeper.nameservices.my-alluxio-cluster=node1,node2,node3
  • alluxio.master.zookeeper.address.[logical-domain]. [Zookeeper node ID] Address foreach Zookeeper node

For each Zookeeper node previously configured, set the full address of each Zookeeper node, for example:

alluxio.master.zookeeper.address.my-alluxio-cluster.node1=host1:2181
alluxio.master.zookeeper.address.my-alluxio-cluster.node2=host2:2181
alluxio.master.zookeeper.address.my-alluxio-cluster.node3=host3:2181

Common Operations

Below are common operations to perform on an Alluxio cluster.

Stop Alluxio

To stop an Alluxio service, run:

$ ./bin/alluxio-stop.sh all

This will stop all the processes on all nodes listed in conf/workers and conf/masters.

You can stop just the masters and just the workers with the following commands:

$ ./bin/alluxio-stop.sh masters # stops all masters in conf/masters
$ ./bin/alluxio-stop.sh workers # stops all workers in conf/workers

If you do not want to use ssh to login to all the nodes and stop all the processes, you can run commands on each node individually to stop each component. For any node, you can stop a master or worker with:

$ ./bin/alluxio-stop.sh master # stops the local master
$ ./bin/alluxio-stop.sh worker # stops the local worker

Restart Alluxio

Starting Alluxio is similar. If conf/workers and conf/masters are both populated, you can start the cluster with:

$ ./bin/alluxio-start.sh all

You can start just the masters and just the workers with the following commands:

$ ./bin/alluxio-start.sh masters # starts all masters in conf/masters
$ ./bin/alluxio-start.sh workers # starts all workers in conf/workers

If you do not want to use ssh to login to all the nodes and start all the processes, you can run commands on each node individually to start each component. For any node, you can start a master or worker with:

$ ./bin/alluxio-start.sh master # starts the local master
$ ./bin/alluxio-start.sh worker # starts the local worker

Add/Remove Workers Dynamically

Adding a worker to an Alluxio cluster dynamically is as simple as starting a new Alluxio worker process, with the appropriate configuration. In most cases, the new worker’s configuration should be the same as all the other workers’ configuration. Run the following command on the new worker to add

$ ./bin/alluxio-start.sh worker SudoMount # starts the local worker

Once the worker is started, it will register itself with the Alluxio leading master and become part of the Alluxio cluster.

Removing a worker is as simple as stopping the worker process.

$ ./bin/alluxio-stop.sh worker # stops the local worker

Once the worker is stopped, and after a timeout on the master (configured by master parameter alluxio.master.worker.timeout), the master will consider the worker as “lost”, and no longer consider it as part of the cluster.

Add/Remove Masters

In order to add a master, the Alluxio cluster must operate in HA mode. If you are running the cluster as a single master cluster, you must configure it to be an HA cluster before having more than one master.

See the journal management documentation for more information about adding and removing masters.

Update Master-side Configuration

In order to update the master-side configuration, you can first stop the service, update the conf/alluxio-site.properties file on master node, and then restart the service. Note that, this approach introduces downtime of the Alluxio service.

Alternatively, one benefit of running Alluxio in HA mode is to use rolling restarts to minimize downtime when updating configurations:

  1. Update the master configuration on all the master nodes without restarting any master.
  2. Restart standby masters one by one (the cluster cannot survive more than floor(n/2) simultaneous restarts).
  3. Elect a standby master as the leading master (tutorial here).
  4. Restart the old leading master that is now a standby master.
  5. Verify the configuration update.

Update Worker-side Configuration

If you only need to update some local configuration for a worker (e.g., change the mount of storage capacity allocated to this worker or update the storage directory), the master node does not need to be stopped and restarted. Simply stop the desired worker, update the configuration (e.g., conf/alluxio-site.properties) file on that node, and then restart the process.