{
  "schema_version": 2,
  "id": "integrate/redis-data-integration/installation/install-vm",
  "title": "Install on VMs",
  "url": "https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/installation/install-vm/",
  "summary": "Learn how to install RDI on one or more VMs",
  "content": "\nThis guide explains how to install Redis Data Integration (RDI) on one or more VMs and integrate it with\nyour source database. You can also\n[Install RDI on Kubernetes](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/installation/install-k8s).\n\n\u003e [!NOTE]\n\u003e We recommend you always use the latest version, which is RDI v.\n\n## Create the RDI database\n\nRDI uses a database on your Redis Enterprise cluster to store its state\ninformation. Use the Redis Enterprise Cluster Manager UI to create the RDI database with the following\nrequirements:\n\n* Redis Enterprise v6.4 or greater for the cluster.\n* For production, 250MB RAM with one primary and one replica is recommended, but for the\n  quickstart or for development, 125MB and a single shard is sufficient.\n* If you are deploying RDI for a production environment then secure this database with a password\n  and TLS.\n* Set the database's\n  [eviction policy](https://redis.io/docs/latest/operate/rs/databases/memory-performance/eviction-policy) to `noeviction`. Note that you can't set this using\n  [`rladmin`](https://redis.io/docs/latest/operate/rs/references/cli-utilities/rladmin),\n  so you must either do it using the admin UI or with the following\n  [REST API](https://redis.io/docs/latest/operate/rs/references/rest-api)\n  command:\n\n  ```bash\n  curl -v -k -d '{\"eviction_policy\": \"noeviction\"}' \\\n    -u '\u003cUSERNAME\u003e:\u003cPASSWORD\u003e' \\\n    -H \"Content-Type: application/json\" \\\n    -X PUT https://\u003cCLUSTER_FQDN\u003e:9443/v1/bdbs/\u003cBDB_UID\u003e\n  ```\n* Set the database's\n  [data persistence](https://redis.io/docs/latest/operate/rs/databases/configure/database-persistence)\n  to AOF - fsync every 1 sec. Note that you can't set this using\n  [`rladmin`](https://redis.io/docs/latest/operate/rs/references/cli-utilities/rladmin),\n  so you must either do it using the admin UI or with the following\n  [REST API](https://redis.io/docs/latest/operate/rs/references/rest-api)\n  commands:\n\n  ```bash\n  curl -v -k -d '{\"data_persistence\":\"aof\"}' \\\n    -u '\u003cUSERNAME\u003e:\u003cPASSWORD\u003e' \\\n    -H \"Content-Type: application/json\" \n    -X PUT https://\u003cCLUSTER_FQDN\u003e:9443/v1/bdbs/\u003cBDB_UID\u003e\n  curl -v -k -d '{\"aof_policy\":\"appendfsync-every-sec\"}' \\\n    -u '\u003cUSERNAME\u003e:\u003cPASSWORD\u003e' \\\n    -H \"Content-Type: application/json\" \\\n    -X PUT https://\u003cCLUSTER_FQDN\u003e:9443/v1/bdbs/\u003cBDB_UID\u003e\n  ```\n If you don't have permissions to use AOF persistence, please check the [Using RDI without persistence](https://redis.io/docs/latest/integrate/redis-data-integration/faq#can-i-use-rdi-without-persistence-enabled) section in the FAQ.\n\n* **Ensure that the RDI database is not clustered.** RDI will not work correctly if the\n  RDI database is clustered (but note that the target database *can* be clustered without\n  any problems).\n\n  If the **Database clustering** option is checked when you create the RDI database (as shown below),\n  you must *uncheck* it before proceeding.\n\n  ![images/rdi/ingest/RDIClusterSetting.webp](https://redis.io/docs/latest/images/rdi/ingest/RDIClusterSetting.webp)\n\n  You can check if your RDI database is clustered from its **Configuration** tab in the\n  Redis Enterprise console. The **Database clustering** option should be set to **None**,\n  as shown in the following screenshot:\n\n  ![images/rdi/ingest/RDICheckUnclustered.webp](https://redis.io/docs/latest/images/rdi/ingest/RDICheckUnclustered.webp)\n\n  If you find the database has been clustered by mistake, you must create a new database with\n  clustering disabled before continuing with the RDI installation.\n\n\n## Hardware sizing\n\nRDI is mainly CPU and network bound. \nEach of the RDI VMs should have at least:\n\n* **CPU**: A minimum of 4 CPU cores. You should consider adding\n  2-6 extra cores on top of this if your dataset is big and you want to ingest the\n  baseline snapshot as fast as possible.\n* **RAM**: 8GB \n* **Disk**: On top of the OS footprint,\n  RDI requires 20GB in the `/var` folder and 1GB in the `/opt` folder (to\n  store the log files). This allows space for upgrades.\n* **Network interface**: 10GB or more.\n\n## VM Installation Requirements\n\nYou would normally install RDI on two VMs for High Availability (HA) but you can also install\njust one VM if you don't need this. For example, you might not need HA during\ndevelopment and testing.\n\n\u003e [!NOTE]\n\u003e You can't install RDI on a host where a Redis Enterprise cluster\n\u003e is also installed, due to incompatible network rules. If you want to install RDI on a\n\u003e host that you have previously used for Redis Enterprise then you must\n\u003e use [`iptables`](https://www.netfilter.org/projects/iptables/index.html) to\n\u003e \"clean\" the host before installation with the following command line:\n\u003e\n\u003e ```bash\n\u003e  sudo iptables-save | awk '/^[*]/ { print $1 } \n\u003e                      /^:[A-Z]+ [^-]/ { print $1 \" ACCEPT\" ; }\n\u003e                      /COMMIT/ { print $0; }' | sudo iptables-restore\n\u003e ```\n\u003e\n\u003e You may encounter problems if you use `iptables` v1.6.1 and earlier in\n\u003e `nftables` mode. Use `iptables` versions later than v1.6.1 or enable the `iptables`\n\u003e legacy mode with the following commands:\n\u003e\n\u003e ```bash\n\u003e sudo update-alternatives --set iptables /usr/sbin/iptables-legacy\n\u003e sudo update-alternatives --set ip6tables /usr/sbin/ip6tables-legacy\n\u003e ```\n\u003e\n\u003e Also, `iptables` versions 1.8.0-1.8.4 have known issues that can prevent RDI\n\u003e from working, especially on RHEL 8. Ideally, use `iptables` v1.8.8, which is\n\u003e known to work correctly with RDI.\n\nThe supported OS versions for RDI are:\n\n* RHEL 8 or 9\n* Ubuntu 20.04, 22.04, or 24.04\n\nYou must run the RDI installer as a privileged user because it installs\n[containerd](https://containerd.io/) and registers services. However, you don't\nneed any special privileges to run RDI processes for normal operation.\n\nRDI has a few\nrequirements for cloud VMs that you must implement before running the\nRDI installer, or else installation will fail. The following sections\ngive full pre-installation instructions for [RHEL](#firewall-rhel) and\n[Ubuntu](#firewall-ubuntu).\n\n### RHEL {#firewall-rhel}\n\nWe recommend you turn off\n[`firewalld`](https://firewalld.org/documentation/)\nbefore installation using the command:\n\n```bash\nsudo systemctl disable firewalld --now\n```\n\nHowever, if you do need to use `firewalld`, you must add the following rules:\n\n```bash\nsudo firewall-cmd --permanent --add-port=443/tcp # RDI API\nsudo firewall-cmd --permanent --add-port=6443/tcp # kube-apiserver\nsudo firewall-cmd --permanent --zone=trusted --add-source=10.42.0.0/16 # Kubernetes pods\nsudo firewall-cmd --permanent --zone=trusted --add-source=10.43.0.0/16 # Kubernetes services\nsudo firewall-cmd --reload\n```\n\nIf you have `nm-cloud-setup.service` enabled, you must disable it and reboot the\nnode with the following commands:\n\n```bash\nsudo systemctl disable nm-cloud-setup.service nm-cloud-setup.timer\nsudo reboot\n```\n\n### Ubuntu {#firewall-ubuntu}\n\nWe recommend you turn off\n[Uncomplicated Firewall](https://wiki.ubuntu.com/UncomplicatedFirewall) (`ufw`)\nbefore installation with the command:\n\n```bash\nsudo ufw disable\n```\n\nHowever, if you do need to use `ufw`, you must add the following rules:\n\n```bash\nsudo ufw allow 443/tcp # RDI API\nsudo ufw allow 6443/tcp # kube-apiserver\nsudo ufw allow from 10.42.0.0/16 to any # Kubernetes pods\nsudo ufw allow from 10.43.0.0/16 to any # Kubernetes services\nsudo ufw reload\n```\n\n## Installation steps\n\nFollow the steps below for each of your VMs.\n\n\u003e [!NOTE]\n\u003e RDI installs executables by default in the `/var` partition, so you must\n\u003e ensure it is mounted without the `noexec` option. Use the following command to\n\u003e find any partitions mounted with the `noexec` option:\n\u003e\n\u003e ```bash\n\u003e mount | grep noexec\n\u003e ```\n\u003e\n\u003e If your `/var` partition is listed in the output from this command, you must remount\n\u003e it without the `noexec` option. See\n\u003e [Using the mount command](https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/5/html/deployment_guide/chap-using_the_mount_command) in the Red Hat documentation to learn how to remount a partition.\n\n1.  Download the RDI installer from the\n    [Redis download center](https://redis-enterprise-software-downloads.s3.amazonaws.com/redis-di/rdi-installation-.tar.gz)\n    (from the *Modules, Tools \u0026 Integration* category) and extract it to your preferred installation\n    folder.\n\n    ```bash\n    export RDI_VERSION=\n    wget https://redis-enterprise-software-downloads.s3.amazonaws.com/redis-di/rdi-installation-$RDI_VERSION.tar.gz\n    tar -xvf rdi-installation-$RDI_VERSION.tar.gz\n    ```\n\n1.  Go to the installation folder:\n\n    ```bash\n    cd rdi_install/$RDI_VERSION\n    ```\n\n1.  Run the `install.sh` script as a privileged user:\n\n    ```bash\n    sudo ./install.sh\n    ```\n\n    \u003e [!NOTE]\n    \u003e RDI uses [K3s](https://k3s.io/) as part of its implementation.\n    \u003e By default, the installer installs K3s in the `/var/lib` directory,\n    \u003e but this might be a problem if you have limited space in `/var`\n    \u003e or your company policy forbids you to install there. You can\n    \u003e select a different directory for the K3s installation using the\n    \u003e `--installation-dir` option with `install.sh`:\n    \u003e\n    \u003e ```bash\n    \u003e sudo ./install.sh --installation-dir \u003ccustom-installation-directory\u003e\n    \u003e ```\n\n    **Advanced**: You can also pass custom K3s parameters to the installer using the\n    `INSTALL_K3S_EXEC` environment variable. For example, to set the kubeconfig file \n    permissions to be readable by all users:\n\n    ```bash\n    sudo INSTALL_K3S_EXEC='--write-kubeconfig-mode=644' ./install.sh\n    ```\n\n    You can combine multiple K3s options in the `INSTALL_K3S_EXEC` variable. See the\n    [K3s documentation](https://docs.k3s.io/installation/configuration) for a full list of \n    available options.\n\n    \u003e [!WARNING]\n    \u003e Only modify K3s parameters if you understand exactly what you are changing \n    \u003e and why. Incorrect K3s configuration can cause RDI installation to fail or result in an \n    \u003e unstable deployment. \n    \n\nThe RDI installer collects all necessary configuration details and alerts you to potential issues, \noffering options to abort, apply fixes, or provide additional information. \nOnce complete, it guides you through creating secrets and setting up your pipeline.\n\n\u003e [!NOTE]\n\u003e It is strongly recommended to specify a hostname rather than an IP address for\n\u003e connecting to your RDI database, for the following reasons:\n\u003e\n\u003e -   Any DNS resolution issues will be detected during the installation rather than\n\u003e     later during pipeline deployment.\n\u003e -   If you use TLS, your RDI database CA certificate must contain the hostname you specified\n\u003e     either as a common name (CN) or as a subject alternative name (SAN). CA certificates\n\u003e     usually don't contain IP addresses.\n\n\u003e [!NOTE]\n\u003e If you specify `localhost` as the address of the RDI database server during\n\u003e installation then the connection will fail if the actual IP address changes for the local\n\u003e VM. For this reason, we recommend that you don't use `localhost` for the address. However,\n\u003e if you do encounter this problem, you can fix it using the following commands on the VM\n\u003e that is running RDI itself:\n\u003e\n\u003e ```bash\n\u003e sudo k3s kubectl delete nodes --all\n\u003e sudo service k3s restart\n\u003e ```\n\nAfter the installation is finished, RDI is ready for use.\n\n### Supply cloud DNS information\n\n\u003e [!NOTE]\n\u003e This section is only relevant if you are installing RDI\n\u003e on VMs in a cloud environment.\n\nIf you are using [Amazon Route 53](https://aws.amazon.com/route53/),\n[Google Cloud DNS](https://cloud.google.com/dns?hl=en), or\n[Azure DNS](https://azure.microsoft.com/en-gb/products/dns)\nthen you must supply the installer with the nameserver IP address\nduring installation. The table below\nshows the appropriate IP address for each cloud provider:\n\n| Platform | Nameserver IP |\n| :-- | :-- |\n| [Amazon Route 53](https://aws.amazon.com/route53/) | 169.254.169.253 |\n| [Google Cloud DNS](https://cloud.google.com/dns?hl=en) | 169.254.169.254 |\n| [Azure DNS](https://azure.microsoft.com/en-gb/products/dns) | 168.63.129.16 |\n\nIf you are using Route 53, you should first check that your VPC\nis configured to allow it. See\n[DNS attributes in your VPC](https://docs.aws.amazon.com/vpc/latest/userguide/AmazonDNS-concepts.html#vpc-dns-support)\nin the Amazon docs for more information.\n\n### Installing with High Availability\n\nTo install RDI with High Availability (HA), perform the [Installation steps](#installation-steps)\non two different VMs. The first VM will automatically become the active (primary) instance, \nwhile the second VM will become the passive (secondary) one. \nWhen starting the RDI installation on the second VM, the installer will detect that the RDI\ndatabase is already in use and ask you to confirm that you intend to install RDI with HA.\n\nAfter the installation is complete, you must set the source and target database secrets\non both VMs as described in [Deploy a pipeline](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/data-pipelines/deploy). If you use `redis-di` to deploy your configuration, you only need to do this on one of the VMs, not both.\n\nIn a High Availability setup, the RDI pipeline is only active on the primary instance (VM).\nThe two RDI instances will use the RDI database for leader election. If the primary instance fails \nto renew the lease in the RDI database, it will lose the leadership and a failover to the secondary instance\nwill take place. After the failover, the secondary instance will become the primary one, \nand the RDI pipeline will be active on that VM.\n\nYou may find it useful to trigger a failover deliberately to check that RDI is correctly configured to handle it. See [Test HA failover](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/installation/ha-test) to learn how to do this.\n\n## Prepare your source database\n\nBefore deploying a pipeline, you must configure your source database to enable CDC. See the\n[Prepare source databases](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/data-pipelines/prepare-dbs)\nsection to learn how to do this.\n\n## Deploy a pipeline\n\nWhen the installation is complete, and you have prepared the source database for CDC,\nyou are ready to start using RDI. See the guides on how to\n[configure](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/data-pipelines) and\n[deploy](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/data-pipelines/deploy)\nRDI pipelines for more information. You can also configure and deploy a pipeline\nusing [Redis Insight](https://redis.io/docs/latest/develop/tools/insight). See\n[RDI in Redis Insight](https://redis.io/docs/latest/develop/tools/insight/rdi-connector)\nfor full details on how to connect to RDI and deploy pipelines.\n\n\u003e [!NOTE]\n\u003e The [`redis-di` CLI](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/reference/cli)\n\u003e is bundled with the VM installer, so it is already available on the VM where RDI is installed. If you\n\u003e prefer to run it from your own laptop or desktop instead, you can\n\u003e [download it separately](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/installation/install-k8s#download-the-rdi-cli)\n\u003e for your platform.\n\n## Configure the Flink processor\n\nRDI ships with two stream processor implementations: the default *classic*\nprocessor and the\n[Apache Flink](https://flink.apache.org/)-based *Flink* processor. \nSee\n[Stream processor implementations](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/architecture#stream-processor-implementations)\nfor an overview of the differences and\n[Differences between the classic and Flink processors](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/architecture/classic-vs-flink)\nfor a side-by-side comparison.\n\nTo run a specific pipeline on the Flink processor, set\n[`processors.type`](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/data-pipelines/pipeline-config#processors)\nto `flink` in that pipeline's `config.yaml` and redeploy it. Pipelines without\nthis setting continue to use the classic processor. Fine-tune the Flink runtime\nthrough the `processors.advanced` section of `config.yaml` (see the\n[configuration reference](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/reference/config-yaml-reference#processors)).\n\nFor migrating existing pipelines to the Flink processor, see\n[Migrate from the classic processor to the Flink processor](https://redis.io/docs/latest/integrate/redis-data-integration/1.19.1/installation/migration-classic-to-flink).\n\n## Uninstall RDI\n\nIf you want to remove your RDI installation, go to the installation folder and run\nthe uninstall script as a privileged user:\n\n```bash\nsudo ./uninstall.sh\n```\n\nThe script will ask if you are sure before proceeding:\n\n```\nThis will uninstall RDI and its dependencies, are you sure? [y, N]\n```\n\nIf you type anything other than \"y\" here, the script will abort without making any changes\nto RDI or your source database.\n",
  "tags": ["docs","integrate","rs","rdi"],
  "last_updated": "2026-09-19T17:55:58-07:00"
}