Riak nodes now negotiate with each other to determine supported operating modes. This allows clusters containing mixed-versions of Riak to properly interoperate without special configuration, and simplifies rolling upgrades.

In previous Riak versions, users were required to disable new features during the rolling upgrade process, and then enable them after all nodes were upgraded. This is now handled automatically by Riak.

Debian/Ubuntu

The following example demonstrates upgrading a Riak node that has been installed with the Debian packages provided by Basho.

1. Stop Riak

riak stop

2. Backup Riak etc and data directories

sudo tar -czf riak_backup.tar.gz /var/lib/riak /etc/riak

3. Upgrade Riak

sudo dpkg -i <riak_package_name>.deb

4. Restart Riak

riak start

5. Verify Riak is running the new version

riak-admin status

6. Wait for the riak_kv service to start

riak-admin wait-for-service riak_kv <target_node>

• <target_node> is the node which you have just upgraded (e.g. riak@192.168.1.11)

7. Wait for any hinted handoff transfers to complete

riak-admin transfers

• While the node was offline other nodes may have accepted writes on it's behalf. This data is transferred to the node when it becomes available.

8. Repeat the process for the remaining nodes in the cluster

RHEL/Centos

The following example demonstrates upgrading a Riak node that has been installed with the RHEL/Centos packages provided by Basho.

1. Stop Riak

riak stop

2. Backup Riak etc and data directories

sudo tar -czf riak_backup.tar.gz /var/lib/riak /etc/riak

3. Upgrade Riak

sudo rpm -Uvh <riak_package_name>.rpm

4. Restart Riak

riak start

5. Verify Riak is running the new version

riak-admin status

6. Wait for the riak_kv service to start

riak-admin wait-for-service riak_kv <target_node>

• <target_node> is the node which you have just upgraded (e.g. riak@192.168.1.11)

7. Wait for any hinted handoff transfers to complete

riak-admin transfers

• While the node was offline other nodes may have accepted writes on it's behalf. This data is transferred to the node when it becomes available.

8. Repeat the process for the remaining nodes in the cluster

9. Once all nodes have been upgraded, add the following additions to the app.config file in /etc/riak on each node. First add to the riak_kv section:

{legacy_keylisting, false},
{mapred_system, pipe},
{vnode_vclocks, true}

and then add to the riak_core section:

{platform_data_dir, "/var/lib/riak"}

10.\ Either run riak stop followed by riak start on all of the nodes in the cluster or use riak attach on each node and execute the following commands:

> application:set_env(riak_kv, legacy_keylisting, false).
> application:set_env(riak_kv, mapred_system, pipe).
> application:set_env(riak_kv, vnode_vclocks, true).

Solaris/OpenSolaris

The following example demonstrates upgrading a Riak node that has been installed with the Solaris/OpenSolaris packages provided by Basho.

1. Stop Riak

riak stop

2. Backup Riak etc and data directories

sudo gtar -czf riak_backup.tar.gz /opt/riak/data /opt/riak/etc

3. Uninstall Riak

sudo pkgrm BASHOriak

4. Install the new version of Riak

sudo pkgadd -d <riak_package_name>.pkg

4. Restart Riak

riak start

5. Verify Riak is running the new version

riak-admin status

6. Wait for the riak_kv service to start

riak-admin wait-for-service riak_kv <target_node>

• <target_node> is the node which you have just upgraded (e.g. riak@192.168.1.11)

7. Wait for any hinted handoff transfers to complete

riak-admin transfers

• While the node was offline other nodes may have accepted writes on it's behalf. This data is transferred to the node when it becomes available.

8. Repeat the process for the remaining nodes in the cluster