Search docs.molo17.com

    System Requirements

    Before deploying Gluesync, ensure your environment meets these requirements for optimal performance.

    How-to guides

    For step-by-step walkthroughs tested on a running Gluesync, see the how-to guides Install Gluesync on Linux and size the host and Free disk space and manage cache and logs.

    Quick reference

    The requirements listed below are specifications where Gluesync is best suited, we recommend using our setup wizards to install Gluesync on Linux and Windows systems. The setup wizards will automatically configure the system according to the requirements listed below, including setting up the necessary resources, configurations and dependencies (yes, Docker as well so you don’t have to worry about it).

    Install Gluesync on a dedicated system

    Gluesync should be installed on a dedicated third system, separate from both the source database and the target system.

    Avoid running Gluesync on machines that already host databases or other resource-intensive or intrusive software (for example, endpoint protection/anti-virus), as this can impact performance and operational stability.

    Linux system requirements at a glance

    Aspect Details

    Supported distributions

    Linux LTS releases (Ubuntu 22.04+, RHEL 9+, Alma/Rocky 9, Debian 12, Fedora, openSUSE, Amazon Linux). Choose Kubernetes nodes or Docker/Podman hosts depending on scale. No upfront containers knowledge required if using the Compose kit which comes with the Gluesync auto-installer.

    Hardware baseline

    • Docker/Podman compose (≤8 pipelines): 8 vCPU, 16 GB RAM. Give /opt/gluesync 250 GB. That is the install and working disk.

    • Kubernetes production with HA (~250k+ RPS): 3 nodes totaling 12 vCPU, 24 GB RAM, 250 GB disk

    • Add storage headroom for Prometheus, backups, and source-agent ArenaCache (local CDC backlog on the agent volume).

    Disk layout (Linux)

    Gluesync runs in Docker. Container layers, temporary files, and logs go under /var/lib/docker.

    If the OS is split across partitions, also keep at least 30 GB free on /var, 50 GB recommended, so Docker has room for the images (about 10 GB after a fresh install, more while an update keeps the old and the new images) and the container logs. That space is extra, not part of the 250 GB on /opt/gluesync. When the server has less than 10 GB of swap, the installer also creates a 10 GB /swapfile on /: count it too when /var is on the root disk. The installer checks these sizes from Gluesync Control 1.4.19.

    Path Size What lives here

    /opt/gluesync

    250 GB

    kit, configs, license, working data

    /var (especially /var/lib/docker and /var/lib/containerd)

    30 GB free (50 GB recommended)

    Docker images and layers, temp, logs

    / (only when the server has less than 10 GB of swap)

    10 GB more

    /swapfile, created by the installer

    Container runtime

    Docker Engine + Compose v2, Podman (rootful or rootless) + Docker Compose v2, or kubelet with CRI‑O/containerd.

    Best suited for

    Production workloads, multi-pipeline deployments, air-gapped or headless installations.

    Windows system requirements at a glance

    Aspect Details

    Supported editions

    Windows Server 2022/2025 is recommended for production workloads, use Windows 10/11 (Pro or Enterprise) for lab/testing workloads. Gluesync runs on Windows-based docker images, which can be run on any Windows machine (and VMs too).

    Hardware baseline

    • 8 vCPU

    • 16 GB RAM (8 GB minimum for POC)

    • 250 GB free disk (SSD recommended)

    • Optional: extra 100 GB if enabling Prometheus/Grafana if you require high retention periods for metrics (e.g. 1 month)

    Container runtime

    Docker engine is required, which comes bundled with our installation kit (Auto-Installer). Docker Desktop (bundles Docker Engine + Compose) is also supported for lab environments. Podman is not supported on Windows.

    Reboot required

    A system reboot is required after installation to ensure all components and services are correctly initialized.

    Best suited for

    Small to mid-size Production workloads, POC/demo environments, lab setups, or small single-pipeline deployments that need a Windows UI (not required for Gluesync).

    If your Windows Server runs Docker behind a corporate proxy or needs to grant non-admin users access to the Docker named pipe, follow the dedicated workflow in the Docker installation guide: Windows Docker Engine: daemon.json proxy + group.

    That section explains how to edit C:\ProgramData\Docker\config\daemon.json, encode proxy credentials (for example DOMAIN%5Cuser), restart Docker, validate with docker ps, and troubleshoot //./pipe/docker_engine access errors. Set this up before installing Gluesync to avoid registry pulls failing behind the proxy.

    Docker Inc. and several upstream components announced end-of-life plans for Windows Server 2019 container support. As these dependencies sunset, Gluesync will phase out Windows Server 2019 compatibility in an upcoming release. Begin planning migrations to Windows Server 2022 (or Linux) to remain on supported ground.

    Network allowlist essentials

    Purpose FQDN / pattern

    Docker Hub images

    https://*.docker.io, https://*.docker.com

    Gluesync backoffice APIs

    https://api.backoffice.molo17.com

    Support logs upload (voucher): support portal API

    https://api.support.molo17.com

    Support logs upload (voucher): storage

    https://*.s3.eu-central-1.amazonaws.com

    Support logs upload with ticket number and e-mail (legacy)

    https://webdav.molo17.com, https://ftp.molo17.com

    Gluesync telemetry / log shipping (optional)

    https://logstash.molo17.com

    Gluesync documentation & updates

    https://docs.molo17.com, https://molo17.com, https://support.molo17.com

    AI features — OpenAI (optional)

    https://api.openai.com

    AI features — Anthropic (optional)

    https://api.anthropic.com

    AI features — Azure OpenAI (optional)

    Your Azure OpenAI resource, for example https://*.openai.azure.com

    If outbound internet is blocked, mirror the container images (docker image save > tar) and publish them to your internal registry, then import them with docker image load. For complete offline installation and update procedures, see Offline installation and updates.

    AI features (LLM providers)

    Optional. Replication does not need a model. AI Studio, Query Studio AI helper, and gluesync.ai do: they call your bring-your-own LLM through Settings → LLM providers.

    Aspect Details

    Cloud APIs

    OpenAI, Anthropic, or Azure OpenAI. Core Hub needs outbound HTTPS to the vendor (see the allowlist above). No extra RAM on the Gluesync host. Use a model that supports tool calling.

    Local models (Ollama)

    Run Ollama (HTTP API, port 11434) on this host or another box. Size RAM/GPU for the model in addition to the Gluesync baseline. From Docker/Kubernetes, Core Hub must not use localhost for Ollama on the host — use host.docker.internal or the host IP.

    Who configures it

    Super admin. An empty vault blocks every AI surface.

    Full install, Docker/Ollama URLs, Azure fields, and connection-test notes: AI models and LLM providers.

    Prerequisites checklist

    Containerized deployment setup

    Gluesync is designed to run efficiently in containerized environments. We recommend using Docker, Podman, or Kubernetes for deployment, depending on your needs and familiarity.

    If you’re not familiar with containerized deployment, start with Gluesync’s Docker Compose deployment strategy—it ships with a handy installer that fully automates the setup, so you don’t need to install a separate container runtime when using the provided Compose kit; nor you are required to know anything about Docker and containers at all.

    Docker/Podman setup

    Requirement Details

    Container runtime & compose

    Gluesync Compose deployments use one of the following: * Docker Engine and Docker Compose * Podman 4.9 or later, rootful or rootless, on Linux, with Docker Compose v2 (podman-compose is not used); rootless also needs Core Hub 2.3.2.0 or later and Chronos 0.3.3 or later

    On Linux, sudo gluesync install --runtime swarm instead runs Gluesync as a Docker Swarm stack on Docker Engine, with no Compose: see Deploy Gluesync on Docker Swarm.

    For a standard installation, use the setup wizard at Get Gluesync; you do not need to install the runtime separately before starting the wizard.

    DIY runtime references

    These are for system administrators who must provision the runtime manually:

    Docker: Prepare Docker manually

    Podman: Gluesync on Podman

    Official references: Docker installation, Podman installation, and Windows Server for Windows containers.

    Gluesync Compose Kit

    Follow the setup wizard. On Linux the installer is gluesync install, started by the docker-installer.sh command. Start Gluesync with the kit’s run.sh helper (macOS) or gluesync run (Windows Server). See Install with the Gluesync kit.

    Verify system requirements

    You can autonomously verify if your system meets all the requirements by running our pre-flight checker tool. It runs the same prerequisite checks as the Gluesync installers, with the same codes and exit codes, and provides a detailed report of any missing requirements. It only reads: it installs nothing and changes no setting.

    • Download and run the pre-flight checker: Gluesync Pre-flight Checker

      • Windows Server: download the signed script gluesync-preflight.ps1 from the latest release and run it from an elevated PowerShell. Add -KitId <kit id> to also check your kit ID.

      • Linux: download gluesync-preflight.sh from the latest release and run it with bash. Add --kit-id <kit id> to also check your kit ID.

    • Follow the instructions in the repository to execute the checks

    • Review the generated report to identify and address any issues

    The pre-flight checker verifies:

    • Windows Server version and edition (Windows Server 2019, 2022 or 2025)

    • System resources (CPU, RAM, disk space)

    • Ports 80 and 443, which Gluesync publishes on the server

    • Required software versions

    • Network connectivity

    • Container runtime configuration

    • Optionally your kit ID: its format and that the kit exists on the MOLO17 download server

    • And other environment-specific requirements

    Kubernetes Setup

    Requirement Details

    Kubernetes Cluster

    A running Kubernetes cluster (e.g., Minikube for testing, or a production-grade cluster for deployment). LTS Linux is recommended for production.

    kubectl CLI

    Install and configure kubectl to interact with your cluster.

    Supported distributions

    Kubernetes (recommended), OpenShift, or an equivalent orchestrator. See production specs above for hardware guidance.

    Installation guides

    Gluesync Kubernetes deployment

    Refer to Gluesync documentation or contact support for deployment guides specific to Kubernetes environments.

    Note

    Gluesync documentation does not cover Kubernetes installation. For setup help, refer to the official Kubernetes documentation.

    Source database requirements

    Requirement Details

    Connection details

    • Username and password

    • Connection string (IP address/port)

    • Table names

    • TLS certificates (if secure authentication required)

    Access permissions

    Database user must have appropriate read permissions

    Target system requirements

    Requirement Details

    Connection details

    • Destination (bucket or database)

    • Connection string (IP address/port)

    • Username and password

    • TLS certificates (if secure authentication required)

    Access permissions

    Write permissions to target location

    We recommend using a database management tool for easier setup and testing. Our QA team regularly uses:

    You can also use Query Studio, the built-in SQL workbench embedded in the Gluesync Control Plane. It lets you inspect schemas, run read-only queries, and validate synchronized datasets without installing any external tool.

    Note: These are suggestions only - use any database tool you’re comfortable with.

    Environment Specifications

    POC / Demo environment

    Component Requirement

    Operating system

    Windows, macOS, or Linux (LTS version recommended)

    Container platform

    Docker, Docker Compose, or Kubernetes

    Hardware

    • 8 GB RAM (minimum)

    • 8 vCPU

    • 100 GB free disk space (Additional space needed for Prometheus monitoring)

    Production environment

    Kubernetes

    Recommended specs for handling ~80,000 rows per second (RPS) on a K8s production grade cluster

    Component Requirement

    Operating system

    Linux (LTS version recommended)

    Container platform

    Kubernetes (recommended), OpenShift, or equivalent orchestrator

    Cluster configuration

    3 Kubernetes nodes for High Availability (HA)

    Total hardware resources

    • 32 GB RAM

    • 16 vCPU

    • 200 GB free disk space (Additional space needed for Prometheus monitoring)

    Docker compose / Podman

    Recommended specs for single pipeline or up to two pipelines (small to medium deployments)

    Component Requirement

    Operating system

    Linux (LTS version recommended)

    Container platform

    Docker Compose, or Podman with Docker Compose v2

    Deployment scale

    Suitable for a single pipeline or up to two concurrent pipelines (small to medium deployments)

    Total hardware resources

    • 16 GB RAM

    • 8 vCPU

    • 200 GB free disk space (Additional space needed for Prometheus monitoring)

    These specifications are baseline recommendations. Actual requirements may vary based on:

    • Data volume

    • Number of concurrent connections

    • Complexity of transformations

    • Monitoring requirements