Rs 5 4 2 April 2019

Redis Enterprise Software (RS) 5.4.2 is now available. This release improves the compatibility of Active-Active Redis (CRDB) with open source Redis, adds SFTP and Mount Points as backup destinations, email alerting and a number of other enhancements and bug fixes.

Overview

Follow these [instructions]({{< relref "/operate/rs/installing-upgrading/upgrading" >}}) for upgrading to RS 5.4.2 from RS 5.0 and above. If you have a version older than 5.0, you must first upgrade to version 5.2 (or at least 5.0).

New features

Active-Active Redis (CRDB) {#activeactive-redis-crdb}

RS 5.4.2 introduces new Active-Active Redis (CRDB) capabilities which improve its compatibility with open source Redis. This simplifies the progress of migration from open source Redis to Active-Active Redis:

1. The string data-type in CRDB is now implicitly and dynamically typed, just like open source Redis.
2. Geospatial data is now supported in CRDB.

These changes required upgrading the internal structures of CRDB and the communication protocol between the CRDB replicas. To take advantage of these updates with existing CRDBs you must [upgrade the CRDB protocol]({{< relref "/operate/rs/installing-upgrading/upgrading" >}}#upgrading-crdbs/).

New backup destinations - SFTP and mount point

The storage options for [periodic database backup]({{< relref "/operate/rs/databases/import-export/schedule-backups" >}}), data export, and data import are extended to include:

• SFTP - a secured and well accepted protocol
• A local mount path that can point to network storage

These new options are in addition to the existing backup locations of FTP, AWS S3 and OpenStack Swift.

Email alerts per database (per team member)

Improved distribution of [email alerts]({{< relref "/operate/rs/security/access-control/create-users" >}}) for cluster or database alerts. The cluster administrator can define, for each team member, the specific databases to receive alerts for, and whether to get the cluster alerts.

Optional client authentication (TLS)

You can now fine tune the [TLS configuration]({{< relref "/operate/rs/security/encryption/tls" >}}), and ease certificates management by excluding client authentication enforcement, so that the database clients, such as applications or other clusters, can connect to your database without authentication.

Node maintenance mode

When you do hardware or operating system maintenance on a server that hosts an Redis Enterprise node, it is important that you move all of the shards on that node to another node to protect the data. Starting with RS 5.4.2 you can use [maintenance mode]({{< relref "/operate/rs/clusters/maintenance-mode" >}}) to handle this process simply and efficiently.

Additional capabilities

• Support for open source Redis version 5.0.4

• Viewer roles were further limited to improve security compliance. The DB Viewer and Cluster Viewer roles cannot view the Redis password of a database. The DB Viewer role also cannot view the ‘log’ screen.

  {{< note >}} These improvements are also included in the REST API and are breaking-changes. If you use the relevant API requests with a DB Viewer or Cluster Viewer role, make sure that you update it accordingly. {{< /note >}}

• External user credentials for periodic backup destinations are now hidden. In REST API responses the password is hashed and in the admin console the password is displayed as asterisks.

Information

• For Redis Enterprise Software 5.4 End of Life (EOL), as well as for previous Redis Enterprise Software versions, see [product lifecycle]({{< relref "/operate/rs/installing-upgrading/product-lifecycle" >}}).

Important fixes

• RS26985 - Fixed a failure of a DB with all-nodes policy, when a node is rebooted
• RS26433 - Fixed a scenario in which the CRDB sync stopped on one of the participating clusters
• RS25968 - Fixed a redis crash on RoF cluster
• RS25835 - Fixed a failure in periodic backup when AOF rewrite is in progress on the same shard
• RS25558 - Enabled updating absolute path in periodic backup FTP setup
• RS25381 - Fixed a race condition while updating a database parameter
• RS24188 - Fixed a failure during database upgrade
• RS23508 - Fixed a failure in ‘rladmin status’ command
• RS22485 - Fixed irregularities in service after increasing the size of a database which uses OSS cluster API

Known limitations

Upgrade

• Before you upgrade a database with RediSearch Module to Redis 5.0, the RediSearch [Module must be upgraded]({{< relref "/operate/oss_and_stack/stack-with-enterprise/install/upgrade-module" >}}) to version 1.4.2 or above.
• Node upgrade fails if the SSL certificates were configured in version 5.0.2 and above by updating the certificates on the disk instead of using the new API. For assistance with this issue, see [SSL/TLS certificate updates]({{< relref "/operate/rs/security/certificates/updating-certificates" >}}) in our documentation or contact Redis support.
• We recommend that you test module upgrade commands in a test environment before you upgrade modules in a production environment. The module upgrade arguments are not validated during the upgrade process and incorrect arguments can cause unexpected downtime.
• The rladmin option to use to preserve the current Redis major.minor version during database upgrade is now keep_redis_version, instead of keep_current_version.

Cluster API

• The API for node removal was updated. It must include json data, and "Content-Type: application/json" header. For example:

  shCopy

  curl -X POST -H "Content-Type: application/json" -i -k -u [email protected]:passsword https://localhost:9443/v1/nodes/3/actions/remove --data "{}"
  

Redis commands

• The capability of disabling specific Redis commands does not work on Redis Module specific commands.
• The CLIENT ID command cannot guarantee incremental IDs between clients that connect to different nodes under multi proxy policies.
• CLIENT UNBLOCK command is not supported in RS 5.4 and RS 5.4.2
• Starting from RS 5.4.2 and after upgrading the CRDB, TYPE command for string data-type in CRDBs return "string" (OSS Redis standard).