7 Clustering is a mechanism that enables multiple processes and programs to work
8 together as one entity. For example, when you search for something on
9 google.com, it may seem like your search request is processed by only one web
10 server. In reality, your search request is processed by may web servers
11 connected in a cluster. Similarly, you can have multiple instances of
12 OpenDaylight working together as one entity.
14 Advantages of clustering are:
16 * Scaling: If you have multiple instances of OpenDaylight running, you can
17 potentially do more work and store more data than you could with only one
18 instance. You can also break up your data into smaller chunks (shards) and
19 either distribute that data across the cluster or perform certain operations
20 on certain members of the cluster.
21 * High Availability: If you have multiple instances of OpenDaylight running and
22 one of them crashes, you will still have the other instances working and
24 * Data Persistence: You will not lose any data stored in OpenDaylight after a
25 manual restart or a crash.
27 The following sections describe how to set up clustering on both individual and
28 multiple OpenDaylight instances.
30 Multiple Node Clustering
31 ------------------------
33 The following sections describe how to set up multiple node clusters in OpenDaylight.
35 Deployment Considerations
36 ^^^^^^^^^^^^^^^^^^^^^^^^^
38 To implement clustering, the deployment considerations are as follows:
40 * To set up a cluster with multiple nodes, we recommend that you use a minimum
41 of three machines. You can set up a cluster with just two nodes. However, if
42 one of the two nodes fail, the cluster will not be operational.
44 .. note:: This is because clustering in OpenDaylight requires a majority of the
45 nodes to be up and one node cannot be a majority of two nodes.
47 * Every device that belongs to a cluster needs to have an identifier.
48 OpenDaylight uses the node's ``role`` for this purpose. After you define the
49 first node's role as *member-1* in the ``akka.conf`` file, OpenDaylight uses
50 *member-1* to identify that node.
52 * Data shards are used to contain all or a certain segment of a OpenDaylight's
53 MD-SAL datastore. For example, one shard can contain all the inventory data
54 while another shard contains all of the topology data.
56 If you do not specify a module in the ``modules.conf`` file and do not specify
57 a shard in ``module-shards.conf``, then (by default) all the data is placed in
58 the default shard (which must also be defined in ``module-shards.conf`` file).
59 Each shard has replicas configured. You can specify the details of where the
60 replicas reside in ``module-shards.conf`` file.
62 * If you have a three node cluster and would like to be able to tolerate any
63 single node crashing, a replica of every defined data shard must be running
64 on all three cluster nodes.
66 .. note:: This is because OpenDaylight's clustering implementation requires a
67 majority of the defined shard replicas to be running in order to
68 function. If you define data shard replicas on two of the cluster nodes
69 and one of those nodes goes down, the corresponding data shards will not
72 * If you have a three node cluster and have defined replicas for a data shard
73 on each of those nodes, that shard will still function even if only two of
74 the cluster nodes are running. Note that if one of those remaining two nodes
75 goes down, the shard will not be operational.
77 * It is recommended that you have multiple seed nodes configured. After a
78 cluster member is started, it sends a message to all of its seed nodes.
79 The cluster member then sends a join command to the first seed node that
80 responds. If none of its seed nodes reply, the cluster member repeats this
81 process until it successfully establishes a connection or it is shut down.
83 * After a node is unreachable, it remains down for configurable period of time
84 (10 seconds, by default). Once a node goes down, you need to restart it so
85 that it can rejoin the cluster. Once a restarted node joins a cluster, it
86 will synchronize with the lead node automatically.
88 Setting Up a Multiple Node Cluster
89 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
91 To run OpenDaylight in a three node cluster, perform the following:
93 First, determine the three machines that will make up the cluster. After that,
94 do the following on each machine:
96 #. Copy the OpenDaylight distribution zip file to the machine.
97 #. Unzip the distribution.
98 #. Open the following .conf files:
100 * configuration/initial/akka.conf
101 * configuration/initial/module-shards.conf
103 #. In each configuration file, make the following changes:
105 Find every instance of the following lines and replace _127.0.0.1_ with the
106 hostname or IP address of the machine on which this file resides and
107 OpenDaylight will run::
110 hostname = "127.0.0.1"
112 .. note:: The value you need to specify will be different for each node in the
115 #. Find the following lines and replace _127.0.0.1_ with the hostname or IP
116 address of any of the machines that will be part of the cluster::
119 seed-nodes = ["akka.tcp://opendaylight-cluster-data@127.0.0.1:2550",
120 <url-to-cluster-member-2>,
121 <url-to-cluster-member-3>]
123 #. Find the following section and specify the role for each member node. Here
124 we assign the first node with the *member-1* role, the second node with the
125 *member-2* role, and the third node with the *member-3* role::
131 .. note:: This step should use a different role on each node.
133 #. Open the configuration/initial/module-shards.conf file and update the
134 replicas so that each shard is replicated to all three nodes::
142 For reference, view a sample config files <<_sample_config_files,below>>.
144 #. Move into the +<karaf-distribution-directory>/bin+ directory.
145 #. Run the following command::
147 JAVA_MAX_MEM=4G JAVA_MAX_PERM_MEM=512m ./karaf
149 #. Enable clustering by running the following command at the Karaf command line::
151 feature:install odl-mdsal-clustering
153 OpenDaylight should now be running in a three node cluster. You can use any of
154 the three member nodes to access the data residing in the datastore.
159 Sample ``akka.conf`` file::
163 mailbox-type = "org.opendaylight.controller.cluster.common.actor.MeteredBoundedMailbox"
164 mailbox-capacity = 1000
165 mailbox-push-timeout-time = 100ms
168 metric-capture-enabled = true
172 loggers = ["akka.event.slf4j.Slf4jLogger"]
176 provider = "akka.cluster.ClusterActorRefProvider"
178 java = "akka.serialization.JavaSerializer"
179 proto = "akka.remote.serialization.ProtobufSerializer"
182 serialization-bindings {
183 "com.google.protobuf.Message" = proto
188 log-remote-lifecycle-events = off
190 hostname = "10.194.189.96"
192 maximum-frame-size = 419430400
193 send-buffer-size = 52428800
194 receive-buffer-size = 52428800
199 seed-nodes = ["akka.tcp://opendaylight-cluster-data@10.194.189.96:2550",
200 "akka.tcp://opendaylight-cluster-data@10.194.189.98:2550",
201 "akka.tcp://opendaylight-cluster-data@10.194.189.101:2550"]
203 auto-down-unreachable-after = 10s
215 mailbox-type = "org.opendaylight.controller.cluster.common.actor.MeteredBoundedMailbox"
216 mailbox-capacity = 1000
217 mailbox-push-timeout-time = 100ms
220 metric-capture-enabled = true
224 loggers = ["akka.event.slf4j.Slf4jLogger"]
227 provider = "akka.cluster.ClusterActorRefProvider"
231 log-remote-lifecycle-events = off
233 hostname = "10.194.189.96"
239 seed-nodes = ["akka.tcp://opendaylight-cluster-rpc@10.194.189.96:2551"]
241 auto-down-unreachable-after = 10s
246 Sample ``module-shards.conf`` file::
306 OpenDaylight includes some scripts to help with the clustering configuration.
310 Scripts are stored in the OpenDaylight distribution/bin folder, and
311 maintained in the distribution project
312 `repository <https://git.opendaylight.org/gerrit/p/integration/distribution>`_
313 in the folder distribution-karaf/src/main/assembly/bin/.
315 Configure Cluster Script
316 ^^^^^^^^^^^^^^^^^^^^^^^^
318 This script is used to configure the cluster parameters (e.g. akka.conf,
319 module-shards.conf) on a member of the controller cluster. The user should
320 restart the node to apply the changes.
324 The script can be used at any time, even before the controller is started
329 bin/configure_cluster.sh <index> <seed_nodes_list>
331 * index: Integer within 1..N, where N is the number of seed nodes. This indicates
332 which controller node (1..N) is configured by the script.
333 * seed_nodes_list: List of seed nodes (IP address), separated by comma or space.
335 The IP address at the provided index should belong to the member executing
336 the script. When running this script on multiple seed nodes, keep the
337 seed_node_list the same, and vary the index from 1 through N.
339 Optionally, shards can be configured in a more granular way by modifying the
340 file "custom_shard_configs.txt" in the same folder as this tool. Please see
341 that file for more details.
345 bin/configure_cluster.sh 2 192.168.0.1 192.168.0.2 192.168.0.3
347 The above command will configure the member 2 (IP address 192.168.0.2) of a
348 cluster made of 192.168.0.1 192.168.0.2 192.168.0.3.
350 Set Persistence Script
351 ^^^^^^^^^^^^^^^^^^^^^^
353 This script is used to enable or disable the config datastore persistence. The
354 default state is enabled but there are cases where persistence may not be
355 required or even desired. The user should restart the node to apply the changes.
359 The script can be used at any time, even before the controller is started
364 bin/set_persistence.sh <on/off>
368 bin/set_persistence.sh off
370 The above command will disable the config datastore persistence.
375 OpenDaylight exposes shard information via MBeans, which can be explored with
376 JConsole, VisualVM, or other JMX clients, or exposed via a REST API using
377 `Jolokia <https://jolokia.org/features-nb.html>`_, provided by the
378 ``odl-jolokia`` Karaf feature. This is convenient, due to a significant focus
379 on REST in OpenDaylight.
381 The basic URI that lists a schema of all available MBeans, but not their
386 To read the information about the shards local to the queried OpenDaylight
387 instance use the following REST calls. For the config datastore::
389 GET /jolokia/read/org.opendaylight.controller:type=DistributedConfigDatastore,Category=ShardManager,name=shard-manager-config
391 For the operational datastore::
393 GET /jolokia/read/org.opendaylight.controller:type=DistributedOperationalDatastore,Category=ShardManager,name=shard-manager-operational
395 The output contains information on shards present on the node::
399 "mbean": "org.opendaylight.controller:Category=ShardManager,name=shard-manager-operational,type=DistributedOperationalDatastore",
404 "member-1-shard-default-operational",
405 "member-1-shard-entity-ownership-operational",
406 "member-1-shard-topology-operational",
407 "member-1-shard-inventory-operational",
408 "member-1-shard-toaster-operational"
411 "MemberName": "member-1"
413 "timestamp": 1483738005,
417 The exact names from the "LocalShards" lists are needed for further
418 exploration, as they will be used as part of the URI to look up detailed info
419 on a particular shard. An example output for the
420 ``member-1-shard-default-operational`` looks like this::
424 "mbean": "org.opendaylight.controller:Category=Shards,name=member-1-shard-default-operational,type=DistributedOperationalDatastore",
428 "ReadWriteTransactionCount": 0,
430 "InMemoryJournalLogSize": 1,
431 "ReplicatedToAllIndex": 4,
432 "Leader": "member-1-shard-default-operational",
434 "RaftState": "Leader",
435 "LastCommittedTransactionTime": "2017-01-06 13:19:00.135",
437 "LastLeadershipChangeTime": "2017-01-06 13:18:37.605",
439 "PeerAddresses": "member-3-shard-default-operational: akka.tcp://opendaylight-cluster-data@192.168.16.3:2550/user/shardmanager-operational/member-3-shard-default-operational, member-2-shard-default-operational: akka.tcp://opendaylight-cluster-data@192.168.16.2:2550/user/shardmanager-operational/member-2-shard-default-operational",
440 "WriteOnlyTransactionCount": 0,
441 "FollowerInitialSyncStatus": false,
444 "timeSinceLastActivity": "00:00:00.320",
448 "id": "member-3-shard-default-operational",
452 "timeSinceLastActivity": "00:00:00.320",
456 "id": "member-2-shard-default-operational",
460 "FailedReadTransactionsCount": 0,
461 "StatRetrievalTime": "810.5 μs",
465 "FailedTransactionsCount": 0,
466 "PendingTxCommitQueueSize": 0,
467 "VotedFor": "member-1-shard-default-operational",
468 "SnapshotCaptureInitiated": false,
469 "CommittedTransactionsCount": 6,
470 "TxCohortCacheSize": 0,
471 "PeerVotingStates": "member-3-shard-default-operational: true, member-2-shard-default-operational: true",
473 "StatRetrievalError": null,
476 "AbortTransactionsCount": 0,
477 "ReadOnlyTransactionCount": 0,
478 "ShardName": "member-1-shard-default-operational",
479 "LeadershipChangeCount": 1,
480 "InMemoryJournalDataSize": 450
482 "timestamp": 1483740350,
486 The output helps identifying shard state (leader/follower, voting/non-voting),
487 peers, follower details if the shard is a leader, and other
490 The Integration team is maintaining a Python based `tool
491 <https://github.com/opendaylight/integration-test/tree/master/tools/clustering/cluster-monitor>`_,
492 that takes advantage of the above MBeans exposed via Jolokia, and the
493 *systemmetrics* project offers a DLUX based UI to display the same
496 Geo-distributed Active/Backup Setup
497 -----------------------------------
499 An OpenDaylight cluster works best when the latency between the nodes is very
500 small, which practically means they should be in the same datacenter. It is
501 however desirable to have the possibility to fail over to a different
502 datacenter, in case all nodes become unreachable. To achieve that, the cluster
503 can be expanded with nodes in a different datacenter, but in a way that
504 doesn't affect latency of the primary nodes. To do that, shards in the backup
505 nodes must be in "non-voting" state.
507 The API to manipulate voting states on shards is defined as RPCs in the
508 `cluster-admin.yang <https://git.opendaylight.org/gerrit/gitweb?p=controller.git;a=blob;f=opendaylight/md-sal/sal-cluster-admin-api/src/main/yang/cluster-admin.yang>`_
509 file in the *controller* project, which is well documented. A summary is
514 Unless otherwise indicated, the below POST requests are to be sent to any
517 To create an active/backup setup with a 6 node cluster (3 active and 3 backup
518 nodes in two locations) there is an RPC to set voting states of all shards on
519 a list of nodes to a given state::
521 POST /restconf/operations/cluster-admin:change-member-voting-states-for-all-shards
523 This RPC needs the list of nodes and the desired voting state as input. For
524 creating the backup nodes, this example input can be used::
528 "member-voting-state": [
530 "member-name": "member-4",
534 "member-name": "member-5",
538 "member-name": "member-6",
545 When an active/backup deployment already exists, with shards on the backup
546 nodes in non-voting state, all that is needed for a fail-over from the active
547 "sub-cluster" to backup "sub-cluster" is to flip the voting state of each
548 shard (on each node, active AND backup). That can be easily achieved with the
549 following RPC call (no parameters needed)::
551 POST /restconf/operations/cluster-admin:flip-member-voting-states-for-all-shards
553 If it's an unplanned outage where the primary voting nodes are down, the
554 "flip" RPC must be sent to a backup non-voting node. In this case there are no
555 shard leaders to carry out the voting changes. However there is a special case
556 whereby if the node that receives the RPC is non-voting and is to be changed
557 to voting and there's no leader, it will apply the voting changes locally and
558 attempt to become the leader. If successful, it persists the voting changes
559 and replicates them to the remaining nodes.
561 When the primary site is fixed and you want to fail back to it, care must be
562 taken when bringing the site back up. Because it was down when the voting
563 states were flipped on the secondary, its persisted database won't contain
564 those changes. If brought back up in that state, the nodes will think they're
565 still voting. If the nodes have connectivity to the secondary site, they
566 should follow the leader in the secondary site and sync with it. However if
567 this does not happen then the primary site may elect its own leader thereby
568 partitioning the 2 clusters, which can lead to undesirable results. Therefore
569 it is recommended to either clean the databases (i.e., ``journal`` and
570 ``snapshots`` directory) on the primary nodes before bringing them back up or
571 restore them from a recent backup of the secondary site (see section
572 :ref:`cluster_backup_restore` below).
574 If is also possible to gracefully remove a node from a cluster, with the
577 POST /restconf/operations/cluster-admin:remove-all-shard-replicas
583 "member-name": "member-1"
587 or just one particular shard::
589 POST /restconf/operations/cluster-admin:remove-shard-replica
595 "shard-name": "default",
596 "member-name": "member-2",
597 "data-store-type": "config"
601 Now that a (potentially dead/unrecoverable) node was removed, another one can
602 be added at runtime, without changing the configuration files of the healthy
603 nodes (requiring reboot)::
605 POST /restconf/operations/cluster-admin:add-replicas-for-all-shards
607 No input required, but this RPC needs to be sent to the new node, to instruct
608 it to replicate all shards from the cluster.
612 While the cluster admin API allows adding and removing shards dynamically,
613 the ``module-shard.conf`` and ``modules.conf`` files are still used on
614 startup to define the initial configuration of shards. Modifications from
615 the use of the API are not stored to those static files, but to the journal.
617 .. _cluster_backup_restore:
619 Backing Up and Restoring the Datastore
620 --------------------------------------
622 The same cluster-admin API that is used above for managing shard voting states
623 has an RPC allowing backup of the datastore in a single node, taking only the
624 file name as a parameter::
626 POST /restconf/operations/cluster-admin:backup-datastore
632 "file-path": "/tmp/datastore_backup"
638 This backup can only be restored if the YANG models of the backed-up data
639 are identical in the backup OpenDaylight instance and restore target
642 To restore the backup on the target node the file needs to be placed into the
643 ``$KARAF_HOME/clustered-datastore-restore`` directory, and then the node
644 restarted. If the directory does not exist (which is quite likely if this is a
645 first-time restore) it needs to be created. On startup, ODL checks if the
646 ``journal`` and ``snapshots`` directories in ``$KARAF_HOME`` are empty, and
647 only then tries to read the contents of the ``clustered-datastore-restore``
648 directory, if it exists. So for a successful restore, those two directories
649 should be empty. The backup file name itself does not matter, and the startup
650 process will delete it after a successful restore.
652 The backup is node independent, so when restoring a 3 node cluster, it is best
653 to restore it on each node for consistency. For example, if restoring on one
654 node only, it can happen that the other two empty nodes form a majority and
655 the cluster comes up with no data.