To run one WordPress Multisite network on several servers, deploy identical code and configuration to every web node, give all nodes one authoritative database state, and make every site’s uploads available through shared storage or a deliberately tested replication system. Put DNS, HTTPS, and traffic routing in front of those nodes, coordinate cron and cache work, and automate changes with WP-CLI. Two independently writable databases or unsynchronised local upload folders will eventually diverge.
Antihost offers managed WordPress hosting with free migration, one-click staging and a 14-day free trial.
WordPress describes Multisite as “a collection of sites that all share the same WordPress installation core files.” The sites can share plugins and themes, but each site still has its own media-upload directory and database tables. That distinction defines what must be synchronised: the application build, network database, per-site files, and runtime services.
As an Amazon Associate I earn from qualifying purchases.
What a Multisite network shares—and what it does not
A Multisite network is one coordinated WordPress application rather than a collection of self-contained installations. Network administration, plugin and theme code, and core files are common. Each subsite has its own rows and tables for content and settings, while network-wide tables hold information such as the site list and network options.
Code and configuration are shared
Every web node should run the same WordPress core version, PHP extensions, plugins, themes, must-use plugins, custom code, wp-config.php network constants, and web-server rewrite rules. Keep custom code in version control and promote one tested build instead of updating nodes by hand. A node with one extra plugin version can produce different page output, PHP errors, or database writes from its peers.
#1 Best Overall
- Dell PowerEdge R730xd 24B SFF 2U Server
- 2x Intel Xeon E5-2690 v4 2.6Ghz 14-Core (28-cores Total)
- 128GB DDR4 RAM – 4x 1.2TB 10K SAS 2.5” 12Gb/s
- Dell H730P mini 2GB 12Gb/s RAID
- 2x 750W PSU - 2x 10Gb SFP+ 2x 1Gb (RJ45) NIC
Uploads remain per-site data
Although the installation is shared, media files are stored in separate paths for the individual sites. A request routed to a node that cannot see another node’s upload directory will return broken images, failed downloads, or incomplete galleries. Shared storage or a tested replication/object-storage design must cover reads, writes, deletes, permissions, and large files.
The database is one coordinated state
Per-site tables coexist with network tables in the same database. Copying only one subsite’s tables is not a safe way to clone or synchronise a network because domains, site IDs, network settings, users, and relationships span the installation. Treat the database as a single logical state store, even if its physical service uses replicas.
Free tools Windows power users keep installed
One-click scans. No signup required.
Runtime state needs ownership rules
Page caches, object caches, cron locks, sessions, image-generation jobs, queue records, and generated files can also differ between nodes. Decide where each piece of state lives and how it is invalidated. A load balancer does not make local cache directories or scheduled jobs consistent by itself.
Choose the topology before you copy anything
“Multiple servers” can mean active traffic on several web nodes, a warm standby, a staging mirror, or a one-time move. Each has different write and failover rules. Write down the goal, acceptable data loss (RPO), recovery time (RTO), and who can promote a replacement before building the system.
| Topology | Database write model | Upload approach | Typical purpose |
|---|---|---|---|
| Active web nodes behind a load balancer | One authoritative writer or a managed design with explicit write ownership and failover | Shared storage or conflict-aware replication visible to every node | High availability and traffic distribution |
| Warm standby | Primary accepts writes; standby is promoted during a documented failover | Continuous or scheduled replication with a measured lag | Disaster recovery with simpler write rules |
| Staging mirror | Production is authoritative; staging receives a controlled copy | Snapshot or export/import, with staging writes isolated | Testing releases without changing production |
| One-time migration | Source is frozen while the target is restored and verified | File copy or archive restore during the maintenance window | Moving the network to new infrastructure |
Do not turn a warm standby into an active writer without a promotion procedure. Likewise, do not assume that two independent writable databases will reconcile WordPress changes safely. If a database service supports replicas, define which endpoint accepts writes, how a failed writer is promoted, how lag is measured, and how split-brain is prevented.
Build identical web nodes
1. Pin the application build
Record the WordPress version, PHP version and extensions, web-server version, plugin and theme versions, must-use plugins, and any system packages needed by image processing or game-media downloads. Build an artifact once and deploy that same artifact to each node. Do not run a plugin update on only the node currently receiving traffic.
Recommended Free Tools
2. Keep configuration consistent
Deploy the same network constants and database connection settings in wp-config.php, including the table-prefix arrangement used by the network. Keep rewrite rules, maximum upload limits, PHP timeouts, and security headers aligned. Environment-specific secrets can differ, but their names, permissions, and effective behavior should be equivalent.
Rank #2
- PROCESSOR & MEMORY: Powered by an Intel Xeon Silver 4112 2.60GHz CPU and 16GB DDR4 RAM for reliable server-grade performance
- STORAGE CAPACITY: Equipped with 32TB total storage via four 8TB 12Gb/s SAS hard drives for high-throughput data handling
- RAID CONTROLLER: Features the PERC H740P RAID controller, enabling advanced data protection and flexible storage configuration
- POWER SUPPLY: Dual 550W redundant power supply units ensure continuous uptime and protection against single power source failure
- FLEXIBLE DEPLOYMENT: Ships with no OS installed, allowing administrators to install their preferred operating system or hypervisor
3. Verify filesystem permissions
The web process must be able to read the code and read/write the relevant media paths without making the entire installation world-writable. Test ownership and permissions on a representative subsite, including a directory containing large game screenshots or video files.
4. Make deployments repeatable
A deployment should be able to create a new node from a clean image, install the exact build, mount uploads, load configuration, run health checks, and join the load balancer. Keep a release identifier so an error can be traced to a specific code and configuration set.
Design the database layer around one source of truth
All web nodes must read a consistent view of network data: posts, users, options, domain mappings, and plugin-generated records. The simplest arrangement is one database service used by every node. A replicated database can also work, but only when its replication mode, read routing, failover, and recovery behavior are understood.
Single writer with read replicas
A common pattern is one writable primary and optional read replicas. Route all writes to the primary. Route reads to replicas only when your application and proxy can tolerate replication lag; administrative screens, login flows, and immediately updated content often need the writer’s current view. Monitor lag and stop routing reads to a replica that falls behind.
Failover and split-brain protection
Document the exact sequence for declaring a primary unavailable, fencing it from writes, promoting a replica, changing connection endpoints, and validating the promoted database. If the old primary can still accept traffic, two writers may create conflicting post IDs, options, or plugin records. Recovery is not complete until the old node is prevented from writing or is rebuilt from the authoritative copy.
Backups are separate from replication
Replication can copy an accidental deletion or a corrupt transaction. Keep independent database backups, retain enough history for your recovery objectives, and test restoring the entire network—not just one site’s tables. A restore test should include network tables, user accounts, mapped domains, and representative subsites.
Synchronise uploads and generated files
Shared filesystem
A shared filesystem gives every node the same media paths immediately. Confirm that it provides the locking, latency, throughput, and permissions your workload needs. Test concurrent uploads, deletion from the media library, thumbnail generation, and a node losing and regaining its mount. A mount that silently becomes read-only can leave WordPress reporting successful requests while files are missing.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteRank #3
Object storage or file replication
Object storage or a file-replication process can work when it is designed for WordPress’s write and delete behavior. Define which node or service is authoritative for a newly uploaded object, how quickly other nodes see it, how deletes propagate, and how conflicts are resolved. Test names containing spaces or non-ASCII characters, files larger than the normal upload limit, and interrupted transfers.
Do not rely on ad-hoc local copies
Copying wp-content/uploads to each node once creates a snapshot, not synchronisation. A subsite that uploads a new image on node A will still serve a 404 when the next request reaches node B. Scheduled copying also creates a window in which media appears inconsistently and can resurrect a file that was deleted elsewhere.
Put DNS, HTTPS, and routing in front
Choose subdomains or subdirectories early
WordPress requires a network addressing model: subsites can use subdomains or subdirectories. The choice can be difficult to reconfigure later, so settle it before production. Check wildcard DNS and web-server behavior if using subdomains; check rewrite rules and existing paths if using subdirectories.
Handle mapped domains deliberately
Native domain mapping is available from WordPress 4.5. Each mapped domain must exist in DNS and resolve to the traffic entry point. Keep the mapping records and web-server configuration identical on every node so a request does not work only on the server that happened to receive it.
Issue certificates for every hostname
Use HTTPS for the primary domain and every mapped domain. Server Name Indication (SNI) allows one endpoint to present certificates for additional domains, but each certificate still needs correct coverage and renewal. Test redirects, mixed-content warnings, REST requests, administration, and uploads over HTTPS after a certificate change.
Health checks must test the application
A health check that only confirms an open TCP port can send users to a node with a failed database connection or an unmounted uploads volume. Check the web process, database connectivity, writable temporary paths, and the presence of a known application response. Remove a node from rotation before maintenance and add it back only after these checks pass.
Coordinate cron, queues, sessions, and caches
Run scheduled jobs once
With several web nodes, every node may attempt to run WordPress scheduled events. Disable per-request pseudo-cron on the front-end nodes if your operations model allows it, then run a controlled cron worker or use a lock and queue mechanism. Ensure the worker can reach the authoritative database and uploads. Duplicate jobs can send repeated notifications, regenerate images concurrently, or publish scheduled game articles more than once.
Rank #4
- WIRED NETWORK USB PRINT SERVER: Connect a single USB 2.0 printer to a wired Ethernet LAN (RJ45); 10Base-T, 100Base-TX auto-sensing to ensure a reliable connection, letting you print from any network computer, across the office or over the Internet
- MANUAL NETWORK SETUP REQUIRED: Configuration via web interface (static IP or DHCP) using LPR queue “LP1"; Not plug-and-play, requires intermediate network knowledge for installation; Access our online FAQs for additional helpful tips and instructions
- USB PRINTER COMPATIBILITY: Works with most USB 2.0 printers using standard drivers; Not compatible with USB hubs, multi-function printers with proprietary drivers, or printers requiring full bi-directional communication
- COMPATIBILITY: The USB to Ethernet print server is USB 2.0 compliant and works with macOS and Windows; It also supports LPR network printing and Bonjour Print Services for broad compatibility; Included software is compatible with Windows only
- PRINT FROM ANYWHERE: Print from any computer connected to the Ethernet; This print server doesn’t require a wired connection to a computer, however it must be connected to your networking device (eg. router or switch) with the included RJ45 network cable
Centralise or invalidate caches
Define where page and object caches live. A local cache must be purged on every node after a post, menu, or network-option change, while a shared cache needs an eviction and failure policy. Test that a newly published page, a changed permalink, and a deleted media file do not remain visible from one node after another node processes the change.
Make sessions and temporary files predictable
If a plugin uses PHP sessions or stores temporary upload fragments locally, a user may lose state when the load balancer sends the next request elsewhere. Prefer a shared session store or stateless authentication where the plugin supports it. Keep temporary directories large enough for imports and video thumbnails, and monitor them for cleanup failures.
Use WP-CLI for repeatable, network-aware changes
WP-CLI provides scriptable database export/import, search-replace, plugin and theme management, and Multisite administration. Run commands against the authoritative environment, capture their output, and review a dry run before changing production.
# Inspect the network database update without changing it
wp core update-db --network --dry-run
# Apply a reviewed network database update
wp core update-db --network
# Export the complete database before a migration
wp db export backup.sql
# Preview URL changes before writing them
wp search-replace 'https://old.example' 'https://new.example' --all-tables --dry-run
# Apply the reviewed replacement, including network tables
wp search-replace 'https://old.example' 'https://new.example' --all-tables
# Flush WordPress caches after a reviewed change
wp cache flushAdapt paths, domains, and database credentials to your environment. Use --url when a command must target a particular subsite, and use network-aware options when the operation is intended for all sites. Store scripts with the application release so another operator can reproduce the change.
Move a Multisite to another server or domain
The migration handbook warns that Multisite moves are more complicated because the database contains multiple references to server names and folder locations. A same-domain move still requires copying the files and database together. A domain or path change requires a planned update of every relevant reference.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →- Freeze or schedule the change. Announce a maintenance window or stop content writes so the database and uploads represent one consistent point in time.
- Back up the complete source. Save the full database,
wp-content, configuration, and web-server rewrite rules. Verify that the backup can be read before shutting down the source. - Provision the target build. Install the exact WordPress, PHP, plugin, theme, must-use-plugin, and server-rule versions. Prepare DNS, certificates, database connectivity, and upload storage without sending public traffic yet.
- Restore database and files together. Import the full network database and restore uploads and custom code. Preserve permissions and the directory layout expected by the network.
- Plan URL substitutions. For a domain or path change, inspect
wp_blogs,wp_site, and each site’shome,siteurl, andfileupload_urlvalues. Include mapped domains and web-server redirects in the plan. - Use a serialization-aware replacement. Do not run a blind SQL replacement across serialized data; changing string lengths without updating serialization metadata can corrupt options and plugin settings. Use WP-CLI search-replace or another tool that understands serialized values, preview the result, and inspect the affected rows.
- Check the network database schema. Run
wp core update-db --network --dry-run. Apply--networkonly after reviewing the proposed update and confirming that the backup is usable. - Test before cutover. Visit the main site and representative subsites. Test administrator login, front-end login, media upload and deletion, permalinks, forms, scheduled events, REST and API calls, search, and every important mapped domain.
- Switch traffic and observe. Change DNS or the load-balancer target, watch HTTP errors, PHP logs, database connections, replication lag, cache behavior, and upload failures. Keep the old environment intact until the validation period ends.
- Retain a rollback path. If errors appear, stop writes or restore the documented traffic route, record which data was created after cutover, and follow the tested recovery procedure rather than making ad-hoc edits on both environments.
Validate synchronization with a practical checklist
- Every node reports the same release identifier and WordPress/plugin/theme versions.
- Each node can connect to the intended database endpoint and cannot accidentally write to an old primary.
- A new image uploaded through one node is visible through every node, and a deletion remains deleted.
- Large files, generated thumbnails, and filenames with spaces or non-ASCII characters survive the storage path.
- Network administration can list all subsites, users, plugins, themes, and domain mappings.
- Representative subsites can log in, publish, edit menus, save settings, and load permalinks.
- REST, XML-RPC where intentionally enabled, forms, webhooks, and API integrations reach the same application state.
- Only the designated cron or queue workers execute scheduled jobs, with observable locks and failures.
- Cache purges reach every cache layer, and a newly changed page is not served stale from another node.
- Certificates, redirects, and DNS records cover the primary network domain and all mapped domains.
- Database backups and upload backups restore into an isolated environment with links and media intact.
- Load-balancer health checks remove a node when its database or uploads are unavailable.
- Operators have written promotion, rollback, and communication steps for a failed node or failed migration.
Common failure modes and their fixes
Images work on one server but 404 on another
The nodes are using separate local upload directories or a broken mount. Compare the upload path and permissions, verify the storage mount from the web process, and upload a new test file while watching both nodes. Fix the storage design before changing database URLs.
Posts or settings appear to revert
Requests are reaching different database writers, or a read replica is behind. Confirm the database endpoint used by each node, fence any unintended writer, inspect replication lag, and route consistency-sensitive reads to the authoritative writer.
A migration breaks a plugin or login
A serialized option may have been damaged by a raw SQL replacement, or only some domain references were changed. Restore the backup in an isolated environment, repeat the replacement with a serialization-aware tool, and inspect plugin options, cookies, redirects, and the affected site’s URL fields.
Scheduled actions run twice
Each web node is running pseudo-cron or workers are sharing no lock. Assign one cron owner or implement a shared lock/queue, then monitor job IDs and execution logs to prove that one event has one owner.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →One node returns redirect loops or certificate warnings
Its mapped-domain list, HTTPS termination settings, or rewrite rules differ from the other nodes. Compare the complete configuration and certificate coverage, correct the node, and keep it out of rotation until the same-domain test passes.
How to compare proposed designs
When choosing between managed services, self-managed replicas, shared filesystems, or object storage, score the design against the actual failure modes rather than the server count.
- Consistency: Is there one database writer, and how are failover and replication lag observed?
- Upload semantics: Are writes, reads, deletes, locks, and large files consistent across nodes?
- Recovery: What are the measured restore time, acceptable data loss, rollback steps, and backup retention?
- Operations: Can a new node be built from code, do plugin/theme updates run once, and are cron and cache tasks observable?
- Network behavior: Do DNS records, TLS certificates, SNI, health checks, sessions, and cache purges cover every mapped domain?
- Cost and lock-in: Include database, storage, bandwidth, managed-service, monitoring, and staffing costs, as well as the effort to leave the platform.
What a safe design looks like
For most networks, the defensible baseline is identical stateless web nodes behind a load balancer, one authoritative database writer with tested backups and failover, and shared or carefully replicated uploads. A controlled deployment promotes the same release to every node. A single cron or queue ownership model prevents duplicate work, while shared cache invalidation prevents stale pages. DNS and certificates cover the primary and mapped domains. WP-CLI scripts make exports, URL replacement, database updates, and cache flushes repeatable and auditable.
That architecture does not remove operational work: storage failures, replication lag, plugin behavior, and domain changes still need monitoring and rehearsed recovery. It does establish a clear source of truth, which is the essential condition for keeping a gaming publication network and its subsites consistent as traffic moves between servers.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
