Key Points
- Drupal 10 loses all security support on December 9, 2026. Acquia ends its Drupal 10 support on the same date.
- Two critical PHP vulnerabilities disclosed in October 2026 add a second deadline: sites on PHP 8.1 or older must upgrade PHP before they can apply the fix.
- You can't skip from Drupal 10 to 12. The realistic options are a Drupal 11 upgrade, extended support as a bridge, or replatforming to Liferay DXP.
- In a move to Liferay DXP, content and data migrate automatically, configuration is recreated, and presentation and custom code are rebuilt.
- A phased migration (Secure, Assess, Pilot, Migrate, then Launch and Optimize) secures the current site first, so nothing is rushed to meet the deadline.
This guide covers three things: what the deadline actually means, three realistic options (upgrade to Drupal 11, extended support as a bridge, or Liferay DXP), and a practical, phased path for teams that choose to move to Liferay DXP.
What happens when Drupal 10 reaches end of life (EOL)?
On December 9, 2026, Drupal's security team will stop issuing advisories, patches and core releases for Drupal 10. Acquia ends its Drupal 10 support on the same date. Drupal 12 is expected to ship that same week (Pantheon, Acquia).
What changes for Drupal 10 sites after December 9, 2026
The main risk is what comes after the deadline. When Drupal 7 reached end of life, HeroDevs found two to six new vulnerabilities a month appearing in widely used contributed modules during the first year, many of them high severity.
How the October 2026 PHP vulnerability affects Drupal sites
In October 2026, researchers at Wiz and Calif disclosed two critical flaws in PHP that had gone undetected for more than a decade. CERT-EU said the flaws could allow unauthenticated remote code execution, meaning an attacker could take control of a server without logging in (BBC News).
The response was unusually severe:
- The UK's NCSC, the US NSA, and agencies in the EU, Australia and Japan advised organizations to apply urgent patches, or take systems offline if they could not.
- Finland temporarily shut down its public government web services, and several Dutch government sites went offline.
- Cloudflare reported early attacks already emerging.
For Drupal teams, this adds work on top of the deadline. Fixes were released only for current PHP versions, so sites on PHP 8.1 or older must first upgrade PHP itself. That means two urgent projects at once.
The broader lesson is about dependency. A self-managed PHP stack makes your team responsible for every layer beneath the website: the language runtime, core, contributed modules, and hosting.
Should you upgrade to Drupal 11 or replatform?
Upgrading to Drupal 11 meets the deadline, but it doesn't fix what makes the deadline painful. The same cycle of major upgrades, module compatibility checks and self-managed runtime security comes back with the next version. Replatforming deals with both.
| Factor | Upgrade to Drupal 11 | Replatform to Liferay DXP |
|---|---|---|
| What it solves | The December 9 deadline | The deadline and the upgrade cycle behind it |
| Next forced upgrade | Another major version later; Drupal releases one every two years | Quarterly Liferay releases with vendor-supported upgrade paths |
| Custom code at each upgrade | Every custom and contributed module rechecked for compatibility | Client Extensions sit outside the core, so upgrades don't break them |
| Runtime security (PHP, server) | Patched by your team or your host | Patched by Liferay on Liferay SaaS or PaaS |
| After the project | The same CMS | Portals, self-service, commerce, AI and personalization on one platform |
| Best fit | Content-only sites that are happy with Drupal | Organizations that need customer, partner or citizen portals, not just a website |
Waiting for Drupal 12 doesn't help; you can't skip from 10 to 12, so it still means a Drupal 11 upgrade first, and sites that wait are unsupported from December 9. Staying on Drupal brings the same upgrade cycle back every two years. Liferay DXP replaces it with routine quarterly releases.
For a side-by-side view of the two platforms, see Liferay vs Acquia.
The practical path has three steps:
1. Patch PHP now
Apply the PHP security fix. Sites on PHP 8.1 or older must upgrade PHP first.
2. Keep Drupal covered
Bridge the December 9 deadline with extended support or a minimal Drupal 11 upgrade. Make that bridge the last money you spend on Drupal.
3. Move to Liferay DXP in planned phases
Use the time the bridge buys to migrate in waves, starting with a pilot site.
Why move from Drupal to Liferay DXP
Liferay DXP replaces a website CMS with a digital experience platform. The same platform runs public sites, authenticated experiences, commerce and AI-assisted content.
Upgrades without rebuilds
Liferay ships quarterly releases. Customizations are built with the Liferay Client Extensions framework, which keeps custom code separate from the platform core, so upgrades don't break it. This removes the module-compatibility scramble Drupal teams face at every major version.
Managed security
On Liferay SaaS or PaaS, Liferay applies platform and runtime security updates, so your team isn't patching the stack itself.
More than a website
Customer and partner portals, self-service, B2B and B2C commerce, forms and workflow, all on one platform with one identity and permissions model.
Content at scale
Structured content, reusable fragments, a visual page builder for marketers, multi-site and multilingual management, including native support for right-to-left languages and those not written in Latin script.
AI in the platform
Liferay AI Hub, available as an add-on, supports content creation, translation and personalized experiences inside the platform.
Open integration
Headless APIs (REST and GraphQL) connect Liferay DXP to CRM, ERP, identity providers and any front end.
Customize freely, without putting the platform at risk
With the Liferay Client Extensions framework, teams build with the tools they already know: React, Vue or Angular for the front end, and any language for back-end services. Extensions talk to Liferay only through its APIs, secured with OAuth 2 and limited to the permissions each one needs. They can't alter or break the platform core. Developers get full freedom, and the platform stays secure and supported.
Not sure which path fits your Drupal estate?
Talk to Liferay's team about your options and what a move to Liferay DXP involves.
Talk to our team
Drupal to Liferay DXP feature mapping
Nearly every standard Drupal capability has a native equivalent in Liferay DXP. Most capabilities that need contributed modules in Drupal are built into Liferay DXP. The tables below cover each functional area. Each row shows the Drupal feature or common module, its native Liferay DXP equivalent, and how it migrates, using one of three methods:
Automated
Scripted import with a batch client extension. Used for content and data: nodes, files, taxonomy, users and translations.
Configured
Set up in the Liferay DXP UI. Used for configuration: content structures, workflow, roles, search and SEO settings.
Rebuilt
Redesigned, usually as fragments or Client Extensions. Used for presentation and custom code: themes, templates, layouts and custom modules.
As a rule of thumb, content and data migrate automatically. Configuration is recreated. Presentation and custom code are rebuilt, and that rebuild is where the platform gains its value.
The mapping targets Liferay CMS, the content management system within Liferay DXP. It is built on Liferay Objects: every content type is a content structure, content and files live in Spaces, and everything is delivered through headless APIs to any site or channel. New migrations should land here rather than on the classic web content tools, which are now in maintenance mode.
Content modeling: Drupal content types to content structures
In Liferay CMS, each Drupal content type becomes a content structure, supporting nested fields and relationships to other structures, and each carrying its own workflow. Content is organized in Spaces and can be reused across any number of sites.
| Drupal | Liferay CMS | How it migrates |
|---|---|---|
| Content type (node bundle) | Content structure | Configured: one structure per content type, keeping field names, each with a unique external reference code (ERC) |
| Node | Content entry in a Space | Automated |
| Node revisions | Entry versions | Automated: latest published version, plus history only if required |
| Paragraphs, field collections | A child content structure linked by a one-to-many relationship, or fragments on content pages | Rebuilt: see Pages, layout and presentation |
| Entity reference to node | Relationship to another content structure | Automated, in a second pass once targets exist |
| Entity reference to term | Categories from global or Space-specific vocabularies | Automated |
| Entity reference to user | Author, or a relationship to the user | Automated |
| Media entities | Files with a file structure (Basic Document, External Video, or a custom file structure) | Automated |
| Custom entities (ECK) and non-editorial data | Custom Objects with relationships | Configured; data loaded in the batch client extension |
| Content moderation per type | Workflow configured inside each content structure | Configured |
| Display modes (view modes) | Display page templates and fragments mapped to structure fields | Rebuilt |
Drupal field types to Liferay CMS fields
| Drupal field type | Liferay CMS content structure field |
|---|---|
| Text (plain) | Text |
| Text (plain, long) | Long Text |
| Text (formatted, long), Body | Rich Text |
| Integer | Integer |
| Decimal, Float | Decimal |
| Boolean | Boolean |
| List (text), List (integer) | Picklist, or Multiselect Picklist for multiple values |
| Date | Date, or Date and Time |
| Date range | Two date fields in a nested group |
| Email, Telephone, Link | Text, with validation |
| Image, File, Media reference | Reference to a file in the Space |
| Address | Nested group of text fields |
| Entity reference (node) | Relationship to another content structure |
| Entity reference (taxonomy) | Category assignment |
| Unlimited cardinality | One-to-many relationship to a child structure (or Multiselect Picklist for simple value lists) |
Pages, layout and presentation
Presentation is rebuilt, not converted, because Twig templates and Layout Builder configuration can't be imported. The upside is that marketers get a visual page builder and reusable components they can manage without developers.
| Drupal | Liferay DXP | How it migrates |
|---|---|---|
| Twig theme (templates, preprocess, libraries) | Theme CSS client extension, plus Style Books for design tokens (colors, typography, spacing) | Rebuilt: tokens extracted from the existing design |
| Page template (page.html.twig, regions) | Master Page Templates | Rebuilt: header, footer and global regions defined once |
| Layout Builder sections and blocks | Content pages built from fragments in the page builder | Rebuilt: each layout becomes a page template |
| Custom blocks (block content) | Fragments, plus editable fragment content | Rebuilt as fragments; block text migrated into them |
| Paragraphs (component-based content) | Fragments on content pages, or a child content structure linked by a relationship | Rebuilt: each Paragraph type becomes a fragment or a related content structure |
| Node templates (node--type.html.twig) | Display Page Templates mapped to structure fields | Rebuilt: one display page per content type |
| Views (lists, grids, feeds) | Collections of content entries, shown with the Collection Display fragment | Configured: filters by structure, category and tag |
| Views exposed filters and facets | Collection Filter fragments; Search facets | Configured |
| Views RSS feed | Collections, or headless APIs for feed consumers | Configured |
| Image styles | Adaptive Media image resolutions | Configured: resolutions generated automatically |
| CKEditor 5 | Built-in rich text editor (CKEditor), extendable through an editor config client extension | Configured |
| Front-end JavaScript (libraries, React or Vue components) | Custom element client extensions | Rebuilt or reused: existing components can often be wrapped as-is |
Media, taxonomy, navigation and search
| Drupal | Liferay CMS | How it migrates |
|---|---|---|
| Media library, managed files | Files in Spaces, organized in folders | Automated, preserving folder structure |
| Media types (image, document, video, remote video) | File structures: Basic Document, External Video, or custom file structures with their own fields | Configured, then automated |
| Private files | Space and folder permissions | Automated, with permissions applied on import |
| Shared media across sites | Spaces connected to multiple sites | Configured |
| Taxonomy vocabularies | Vocabularies, global or Space-specific | Automated |
| Taxonomy terms and hierarchy | Categories, nested to any depth | Automated, parents before children |
| Free tagging | Tags | Automated |
| Menus (main, footer, utility) | Navigation Menus | Automated or configured, depending on size |
| Breadcrumbs | Breadcrumb fragment | Configured |
| Search API with Solr or Elasticsearch | Built-in search (Elasticsearch or OpenSearch) | Configured: content is indexed automatically on import |
| Search facets, boosting, synonyms | Search facets, Search Blueprints (boosting, filtering, ranking rules), synonym sets | Configured |
| Search pages | Search page with search bar, results and facet fragments | Configured |
Forms, workflow, publishing and multilingual
| Drupal | Liferay CMS and Objects | How it migrates |
|---|---|---|
| Webform (forms, conditions, multi-page) | Forms built on Objects with form fragments on content pages | Rebuilt per form; historical submissions imported as Object entries |
| Webform handlers (email, remote post) | Object actions (email, webhook, client extension) and notifications | Configured |
| Structured data capture (custom entities + forms) | Custom Objects with generated forms, APIs and admin views | Configured: no code |
| Content Moderation and Workflows (draft, review, published) | Workflow configured in each content structure, with multi-step approvals and role-based assignment | Configured: states and transitions recreated |
| Workspaces (stage a set of changes) | Drafts reviewed through workflow before publishing (Publications isn't available in Liferay CMS) | Configured |
| Scheduler module (publish or unpublish on a date) | Workflow, including timers | Configured |
| Revisions and revert | Entry versions with history | Automated |
| Content Translation | Integrated translation tools; every structure field is localizable | Automated: each translation imported as a locale of the same entry |
| Language negotiation (URL prefix, domain) | Locale prefix in URLs, per-site default language, virtual hosts per site | Configured |
| Right-to-left language support | Native RTL support in themes and the editor | Configured |
| Translation workflow (TMGMT) | Built-in translation tools: XLIFF export and import for translation agencies, machine translation, and AI Hub agents that translate content and send it through a review workflow before publishing | Configured |
Users, access, multisite, SEO and integrations
| Drupal | Liferay DXP | How it migrates |
|---|---|---|
| Users and user fields | Users with custom fields | Automated |
| Roles and permissions | Regular, site, organization and account roles, with fine-grained permissions per asset | Configured: roles recreated, then assigned automatically on user import |
| Group module, Organic Groups | Organizations, user groups and sites with membership | Configured, then automated |
| Profiles (Profile module) | User custom fields; account management | Automated |
| Passwords | Not portable: Drupal's password hashes don't carry over | Plan SSO, or a password-reset email at launch |
| SAML, OpenID Connect, LDAP modules | SAML, OpenID Connect and LDAP built in | Configured |
| Multisite, Domain Access | Many sites on one instance, each with its own domain, theme, language and permissions | Configured: each Drupal site becomes a Liferay site |
| Site templates (install profiles, distributions) | Site Templates; site initializer client extension | Rebuilt |
| Pathauto (URL patterns) | Friendly URLs on every content and display page | Automated: each article's existing alias imported as its friendly URL |
| Redirect module | Redirection tool, with 404 tracking | Automated or configured: Drupal redirects plus old-to-new URL map |
| Metatag (title, description, Open Graph) | SEO settings per content and display page: title, description, canonical, Open Graph, robots | Automated where stored on content; configured on templates |
| Simple XML Sitemap | Built-in XML sitemap | Configured |
| Hreflang | Generated automatically for localized pages | Configured |
| JSON:API, RESTful Web Services, GraphQL module | Headless REST and GraphQL APIs for all content, documents, users and Objects (OpenAPI-documented) | Configured: consumers update endpoints |
| OAuth (Simple OAuth) | OAuth 2 administration | Configured |
| Cron and queues | Scheduler, plus Batch Engine for bulk jobs | Configured |
| Analytics integration (Google Tag Manager, analytics) | Built-in Google Analytics integration: add your measurement ID in site settings, with no module or code needed. Google Tag Manager and other tags can be added the same way. Liferay Data Platform adds content performance, segmentation and personalization insights. | Configured |
| Personalization modules | Segments, personalized experiences and A/B testing, built in | Configured: a new capability, not migrated |
Drupal contributed modules and their Liferay DXP equivalents
A typical enterprise Drupal site depends on dozens of contributed modules, and each one must be compatible with the next major version. In Liferay DXP most of these are core platform features, so the dependency list shrinks.
| Drupal contributed module | Liferay DXP equivalent |
|---|---|
| Paragraphs | Related content structures; fragments |
| Pathauto, Redirect, Metatag, Simple XML Sitemap | Friendly URLs, Redirection, SEO settings, sitemap (all core) |
| Webform | Forms built on Objects |
| Search API, Search API Solr, Facets | Built-in search, facets, Search Blueprints |
| Content Moderation, Workbench, Scheduler | Workflow in each content structure |
| TMGMT | Liferay CMS translation: XLIFF export and import, machine translation, and Liferay AI Hub translation agents |
| Group, Domain Access | Sites, Spaces, organizations, user groups |
| Entity Construction Kit (ECK) | Objects |
| Media, Entity Browser | Files in Spaces, with file structures |
| Views Bulk Operations | Bulk actions in Liferay CMS |
| Commerce | Liferay Commerce (catalogs, pricing, accounts, orders) |
| Feeds, Migrate Plus | Batch client extensions |
| Captcha, reCAPTCHA | CAPTCHA and reCAPTCHA settings |
| Login and SSO modules (samlauth, openid_connect) | SAML and OpenID Connect |
| Custom modules | Client Extensions (see Rebuilding custom Drupal modules as Client Extensions) |
Drupal end-of-life migration roadmap
The safest migration secures the current site first and then moves in waves, so nothing is rushed to meet the deadline.
The five phases run in order: Secure, Assess, Pilot, Migrate, then Launch and Optimize. Phase 1 starts immediately; a go decision follows Assess, and a pilot sign-off follows Pilot.
The five migration phases
| Phase | Key activities | Deliverables | Exit criteria |
|---|---|---|---|
| 1. Secure | Patch PHP; upgrade PHP 8.1 or older; check for compromise; bridge to Drupal 11 or arrange extended support | Patched, supported Drupal estate | No unsupported software in production after December 9 |
| 2. Assess | Inventory sites, content types, fields, node counts, languages, modules, integrations, users and roles; export URL aliases and traffic; identify content to archive | Migration inventory; content model map (field by field); integration list; target architecture; effort estimate | Go decision on scope, waves and budget |
| 3. Pilot | Set up Liferay DXP environments; create Style Book, master page and fragments; build structures for the pilot's content types; run the pipeline on one site; configure SSO | One representative site running on Liferay DXP with real content | Content is accurate, pages approved by marketing, pipeline repeatable |
| 4. Migrate | Build remaining structures, templates and fragments; rebuild custom features as Client Extensions; run the pipeline site by site; build redirect maps; train editors | All sites migrated in staging, editors trained | Validation passed for each wave (see Testing, validation and cutover) |
| 5. Launch and optimize | Content freeze and delta import; DNS cutover; monitor 404s and search console; decommission Drupal; switch on personalization, analytics and Liferay AI Hub | Live sites; Drupal retired | Traffic and rankings stable at the 30-day check |
The pilot is the most important phase. It proves the content model and the import pipeline on real data, so every later wave reuses the same scripts and templates instead of starting over.
The content migration pipeline
Content moves through a scripted, repeatable pipeline rather than by copy and paste. The same scripts run for the pilot, each wave, and the final delta at cutover.
Extract
Use Drupal's core JSON:API to export nodes with their fields, translations and references, or query the database directly for very large volumes. Copy the files directory alongside. Export URL aliases, redirects, menus, taxonomy, users and roles.
Transform
Convert each item into the JSON shape of its target Liferay DXP structure, apply the transformation rules below, and record every item in the ID map.
Load
Package the transformed data as a batch client extension. The transform step writes its output as batch data files, one per entity type, and the extension bundles them with its configuration. When the extension is deployed, Liferay DXP imports the files through the Batch Engine. Packaging the load this way ensures it's repeatable and versioned: the same extension is deployed to the pilot, to each wave and, with the delta files, at cutover, and it can be run against any environment in the same way. Order the batch files so every reference has a target before it is used:
- Configuration: sites, languages, roles, organizations, user groups, structures, templates and fragments (built in the pilot)
- Users, with role and group assignments
- Vocabularies, then categories (parent-first), then tags
- Spaces and folders, then files
- Content entries in the source language
- Translations, as locale versions of the same items
- Second pass: entity references and internal links rewritten from the ID map
- Object data and historical form submissions
- Navigation menus
- Redirects for every old URL that changed
Validate
Compare counts and samples against the source and crawl for broken links. When something fails, fix the transform rule and re-run; never fix migrated content by hand, or the next run will overwrite it.
Content transformation rules
The transform step is where most of the migration's quality is won or lost. These rules cover the cases that result in the most rework when skipped.
| Drupal content | Rule | Why it matters |
|---|---|---|
| Body HTML | Clean before import: strip Drupal-only markup (inline styles, empty paragraphs, legacy shortcodes and text-format tokens); keep semantic HTML | Content then follows the new Style Book instead of old inline formatting |
| Inline images and file links | Find every src and href pointing to /sites/default/files; upload the file to the Space; rewrite the link to the new document URL | Prevents broken images after Drupal is switched off |
| Internal links (/node/123 and aliases) | Rewrite with the ID map to the new Liferay DXP friendly URL | Avoids redirect chains and preserves link equity |
| Entity references | Second pass after all items exist: look up each target in the ID map, then set the reference | Targets must exist before they can be linked |
| Paragraphs | Data-driven Paragraphs (FAQ items, specs, team members) → entries of a child content structure, linked to the parent by a one-to-many relationship. Layout-driven Paragraphs (hero, columns, carousel) → fragments on content pages | Keeps structured data structured and gives marketers layout control |
| Translations | Import the source language first, then each translation as a locale version of the same item, not a separate entry | Keeps language switching and hreflang working |
| Taxonomy | Import vocabularies, then categories parent-first; map term IDs | Hierarchy and filters keep working |
| Authors and dates | Carry over created and modified dates, author and display date | Preserves editorial history and "published on" dates |
| URL aliases | Set each entry's friendly URL to its Drupal alias where possible; otherwise add a 301 redirect | Protects search rankings |
| Metatags | Map title, description and Open Graph fields into SEO fields | No drop in search snippets |
Keep an ID map throughout: a single table of Drupal ID, Liferay ID, old URL and new URL. Every link rewrite, reference and redirect is generated from it.
Rebuilding custom Drupal modules as Client Extensions
Custom Drupal modules are the main reason Drupal upgrades are costly, because every hook and API they rely on can change between major versions. In Liferay DXP, custom functionality is built with the Client Extensions framework. Extensions run alongside the platform and talk to it through APIs and events, so platform upgrades don't break them.
Start by sorting each custom module into one of three groups: no longer needed, replaced by configuration, or rebuilt as an extension. Many custom modules exist only to fill a gap that Liferay DXP covers natively.
| What the Drupal custom module does | Client extension type |
|---|---|
| Theme overrides, global CSS | Theme CSS, Global CSS |
| Site-wide scripts (tracking, widgets) | Global JS |
| Interactive UI component (calculator, configurator, dashboard) | Custom Element (React, Vue, Angular or plain JS) |
| Embedded external app | IFrame |
| Hooks on save or update (hook_node_presave, hook_entity_insert) | Object Action or Workflow Action, calling a microservice |
| Custom business logic, integrations, scheduled sync | Microservice client extension using OAuth to call Liferay DXP APIs |
| Custom notification channels | Notification type |
| Bulk data loading | Batch client extension |
| Site setup scripted (install profile) | Site Initializer |
| Custom editor toolbar | Editor config contributor |
Because extensions use standard web technologies and Liferay DXP's APIs, front-end and integration developers can build them without platform-specific expertise. Existing React or Vue components from the Drupal site can often be reused as Custom Elements.
Testing, validation and cutover
Each wave is validated against the source before it goes live.
Validation checks
- Item counts per content type, language, vocabulary and Space folder match the source (minus content deliberately archived).
- Field-level spot checks on a sample of every content type, including references and dates.
- No broken links or images: crawl the staging site and confirm zero links to old Drupal file paths.
- Every old URL from the inventory returns either the same page or a 301 to the right page.
- Roles and permissions tested with a real user from each role, including restricted content.
- Forms submit correctly and notifications are delivered.
- Accessibility and performance tested on key templates.
- Business owners approve each site.
Cutover runbook
- 1
Before cutover
Agree on a content freeze window with editors.
- 2
Before cutover
Run a final delta import of anything created or changed since the last full run (using Drupal's changed timestamp).
- 3
Cutover day
Apply redirect maps and confirm the sitemap.
- 4
Cutover day
Switch DNS or the CDN origin to Liferay DXP; keep Drupal read-only as a fallback.
- 5
Cutover day
Run smoke tests on top pages, forms, login and search.
- 6
Weeks 1 to 2
Submit the new sitemap, and use Liferay SEO & AEO Studio's scheduled scans and dashboard to monitor 404s, redirect chains, traffic and AI crawler activity daily for two weeks.
- 7
Day 30
Decommission Drupal once the 30-day review confirms stable traffic and rankings, and archive a final database backup.
Migration team and effort drivers
| Role | Responsibility |
|---|---|
| Project lead | Scope, waves, stakeholders, cutover decisions |
| Liferay solution architect | Target architecture, content model, integration design |
| Migration engineer | Extract and transform scripts, API loading, ID map, delta runs |
| Front-end developer | Style Book, master pages, fragments, Custom Element extensions |
| Integration developer | Microservice and action extensions, SSO, third-party connectors |
| Content lead and editors | Content audit, archiving decisions, review and sign-off |
| SEO and analytics lead | URL inventory, redirect maps, tagging, post-launch monitoring |
Effort depends far more on complexity than on page count. The main drivers:
- Number of distinct content types and paragraph types, since each needs a structure and template or fragment.
- Number of custom modules and integrations that must be rebuilt.
- Number of languages and sites.
- Volume of user accounts and the authentication approach.
- How much content is archived rather than migrated.
A site with tens of thousands of articles but five content types can migrate faster than a small site with forty content types and heavy custom code, because bulk content is scripted while presentation and custom features are not.
How to speed up a Drupal migration with scripts and AI
Most migration time goes into five places: discovering what the Drupal estate holds, modeling content, rebuilding the design, rewriting custom code, and checking the result. Each one can be shortened with scripts, purpose-built tools or AI, as long as people still review what the tools produce.
| Where the time goes | Accelerator | What it speeds up |
|---|---|---|
| Discovery and inventory | An audit script that reads Drupal's configuration export and JSON:API, and lists every content type, field, module, language and node count | Turns weeks of manual inventory into a generated report and a first draft of the field mapping |
| Content modeling | AI-assisted mapping: an LLM is given the Drupal field definitions and the mapping tables in this guide, and drafts Liferay DXP content structures for an architect to approve | The architect reviews and corrects a draft instead of writing every structure from scratch |
| Design and theme | An AI design-conversion agent that reads the existing Drupal site and converts its design into Liferay's Client Extensions | Produces the Style Book, design tokens and starting fragments from the live Drupal site, so the front-end team starts from the brand rather than a blank theme |
| Bulk content | Reusable extract and transform scripts, plus a batch client extension template, with one mapping file per content type | The same pipeline runs for every site and every wave; a new site only needs its mapping file |
| Content clean-up | AI for the long tail: cleaning messy body HTML, writing missing alt text, summaries and meta descriptions, and filling translation gaps through Liferay AI Hub | Fixes thousands of pages that would otherwise be archived or left with poor quality |
| Custom modules | AI coding assistants that read each Drupal module, sort it into retire, configure or rebuild, and draft a Client Extension scaffold for the rebuilds | Shortens the custom-code assessment and gives developers a working starting point |
| Validation | Scripts that reconcile item counts and crawl old and new URLs, plus Liferay SEO & AEO Studio scans of staging | Every wave is checked the same way, in minutes rather than days |
Using AI agents with Liferay's MCP server
Liferay DXP can act as a Model Context Protocol (MCP) server, so AI agents built in Liferay AI Hub, or connected LLMs, can read and work with Liferay DXP content through a standard interface. In a migration this supports agents that compare migrated entries with the Drupal source and flag differences, suggest categories and tags for untagged content, or prepare draft entries for an editor to review. Agents act with the permissions of the user who runs them, so they can only see and change what that user can.
Guardrails for AI in a migration
People approve, AI drafts
Every AI-generated structure, Client Extension and content rewrite is reviewed before it reaches production.
Staging first
AI tools and agents run against the staging environment, never directly against the live site.
The ID map stays deterministic
IDs, URLs and redirects come from the ID map and scripts, never from an AI guess.
Approved models only
Customer content goes only to LLMs the organization has approved, connected and governed through Liferay AI Hub.
Everything is logged
Keep a record of what each tool or agent changed, so any run can be checked or repeated.
Build once, reuse for every wave
The biggest saving comes from treating the accelerators as one reusable kit rather than one-off work: the audit script, the mapping workbook, the structure templates, the batch client extension, and the validation suite. The pilot builds the kit; every later wave, and every later Drupal site, reuses it. Implementation partners can carry the same kit from one customer to the next.
How to protect SEO, AEO and users during migration
A migration only succeeds if search rankings, content and user access survive the cutover.
URLs
Export every Drupal path alias, then map each one to its new Liferay DXP URL with a 301 redirect. Keep URL structures identical where possible.
Metadata
Carry over page titles, meta descriptions, canonical tags, Open Graph data and hreflang for multilingual pages.
Sitemaps
Submit the new XML sitemap to search consoles on launch day, and watch crawl errors for the first 30 days.
Content quality
Treat the migration as a clean-up. Archive outdated pages instead of migrating everything.
Users
Migrate accounts and roles, and connect SSO so users don't have to re-register. Plan a password-reset communication if passwords can't be carried over.
Analytics
Re-tag analytics before launch and capture a baseline of traffic and conversions to compare against afterward.
Liferay SEO & AEO Studio
Search has split in two. Buyers still use Google, but they increasingly ask AI assistants such as ChatGPT, Gemini and Perplexity, which return one answer instead of ten links. Answer engine optimization (AEO) is the practice of making content easy for those assistants to find, understand and cite.
Liferay SEO & AEO Studio covers both from inside Liferay DXP. It is an add-on to Liferay AI Hub: it crawls your domains, shows what is holding pages back, and puts AI agents to work fixing it. Because it writes changes back into Liferay CMS, one fix carries to every channel the content is published to. A traditional SEO tool produces a report that turns into developer tickets; the Studio turns most findings into a one-click fix that a marketer approves.
Liferay SEO & AEO Studio
Protect your rankings through the move: scan staging before cutover, fix what the migration broke, and track AI crawler access and citations after launch.
See SEO & AEO Studio
| Capability | What it does | Why it matters in a migration |
|---|---|---|
| Site-wide crawl and insights | Scans metadata, headings, images, URLs, structured data, performance, site structure and AI readiness; groups findings by category and impact | Scan staging before cutover to catch anything the migration broke |
| AI agents that apply the fix | Draft missing meta descriptions, shorten overlong titles, generate schema, and write the summary and FAQ sections answer engines look for | Clears the long tail of migrated pages without a developer backlog |
| Technical fixes | Finds and fixes internal 404s, redirect chains, orphan pages, sitemap errors and AI crawler access | These are exactly the issues a URL change creates |
| Performance dashboard | Search Console clicks, impressions, CTR and position next to PageSpeed metrics, filtered by site, date, country or device | Compares pre- and post-migration performance in one view |
| AI bot monitoring | Shows which AI crawlers (GPTBot, ClaudeBot, PerplexityBot and others) reach your content, and how often | Confirms AI engines can still reach the new site |
| AI citation monitoring | Tracks whether AI assistants cite your pages for the questions buyers ask, month over month | Makes AEO measurable, which classic rank trackers can't do |
| Content structure | Maps page depth, recommends internal links and shows topic clusters and gaps | Rebuilds internal linking around the new information architecture |
| Multi-domain scans | Connects every domain with its own scan schedule, crawler settings and scope | Each migrated site is added as it goes live |
Using SEO & AEO Studio during the migration
- Before cutover, connect the staging site and run a full scan. Fix missing metadata, broken links and schema gaps there, not after launch.
- Capture the Drupal site's Search Console baseline so the dashboard can compare before and after.
- After launch, schedule regular scans of the live domain and watch 404s, redirect chains and AI crawler activity for the first 30 days.
- Once stable, use the AI agents and citation tracking to improve AEO across the whole site, a step Drupal sites usually can't take without extra tools.
Liferay CMS gives AEO a head start: content is structured by design, so each structure field can be exposed as clean, machine-readable data instead of being buried in page markup.
Drupal 10 end of life checklist
Start with this list, whichever option you choose.
Immediate: the PHP vulnerability
Why this step? Your Drupal site stays live on PHP until cutover, which can be weeks or months away, and this vulnerability has been actively exploited. An unpatched site could expose the database, including the users and content you're about to migrate. PHP may also be running in places the migration won't touch, such as other sites, hosting panels and third-party tools.
- Find every server, host and third-party product running PHP, not just the main website
- Upgrade anything on PHP 8.1 or older to a version the fix supports
- Apply the PHP security fix
- Check logs for signs an attacker gained access before patching
Drupal 10 end of life (December 9, 2026)
- List every Drupal site, its version and its owner
- Inventory contributed and custom modules, and check their Drupal 11 compatibility
- Decide for each site: upgrade, replatform, or retire
- Arrange bridge coverage (a Drupal 11 upgrade or extended support) for sites migrating after the deadline
Replatforming to Liferay DXP
- Audit content types, page count, languages and integrations
- Identify the business capabilities you need beyond a website (portals, commerce, self-service)
- Pick a pilot site
- Book a Liferay DXP migration assessment
Drupal 10 end of life FAQ
When is Drupal 10 end of life?
Drupal 10 loses all security support on December 9, 2026. From that date, Drupal's security team stops issuing advisories, patches and core releases for Drupal 10, and Acquia ends its Drupal 10 support. Drupal 12 is expected to ship the same week.
Will my Drupal 10 site stop working after December 9, 2026?
No. The site keeps running, but security patches for Drupal core stop and contributed module maintainers move to Drupal 11 and 12. Running unsupported software can create compliance exposure under PCI DSS, GDPR and regional data-protection rules.
Can I skip Drupal 11 and wait for Drupal 12?
No. You can't upgrade directly from Drupal 10 to 12, so waiting still means a Drupal 11 upgrade first, and sites that wait are unsupported from December 9, 2026.
Does the October 2026 PHP vulnerability affect Drupal sites?
Yes. Drupal runs on PHP, and fixes were released only for current PHP versions. Sites on PHP 8.1 or older must upgrade PHP itself before they can apply the fix.
What migrates automatically from Drupal to Liferay DXP?
Content and data migrate automatically through a scripted import packaged as a batch client extension. Configuration is recreated in the Liferay DXP UI. Presentation and custom code are rebuilt, usually as fragments or Client Extensions.
Can Drupal Twig templates and Layout Builder pages be imported?
No. Twig templates and Layout Builder configuration can't be imported, so presentation is rebuilt with Style Books, master page templates and fragments in the Liferay DXP page builder.
Do Drupal user passwords carry over?
No. Drupal's password hashes don't carry over. Plan SSO, or send a password-reset email at launch.
What happens to custom Drupal modules?
Each one is sorted into one of three groups: no longer needed, replaced by configuration, or rebuilt as a Client Extension. Existing React or Vue components can often be reused as Custom Elements.
How do you protect search rankings during the migration?
Map every Drupal path alias to its new URL with a 301 redirect. Carry over titles, meta descriptions, canonical tags, Open Graph data and hreflang. Submit the new XML sitemap on launch day and watch crawl errors for 30 days.
What drives the effort of a Drupal to Liferay DXP migration?
Complexity more than page count: the number of content and paragraph types, custom modules and integrations, languages and sites, user accounts and authentication approach, and how much content is archived.
Next steps
The deadline is fixed, but the choice is still open. Use this time to secure what you have and decide where you're going, rather than rushing into another two-year upgrade cycle.
Liferay's team can help you:
- Run a free migration assessment of your Drupal estate
- Map your content model to Liferay DXP and estimate the effort
- Plan a pilot so you can see your own content running on Liferay DXP before committing
Plan your move off Drupal 10
Talk to Liferay's team about your Drupal estate and the right next step.
Talk to our team