CentOS Is Dead. Here’s What to Migrate To — and Which VPS Providers Still Offer It (and Why They Shouldn’t)

The CentOS Death Timeline: What Actually Happened

I need to start here because I still get emails from people running CentOS 7 servers in production who genuinely do not know their system is end-of-life. So let me lay this out clearly.

In December 2020, Red Hat dropped a bomb: CentOS Linux — the free, stable, downstream rebuild of RHEL that half the internet ran on — was being killed. Not sunset gradually. Killed. CentOS 8, which had only been released 13 months earlier, would reach end-of-life on December 31, 2021. People who had just migrated from CentOS 7 to CentOS 8 were blindsided.

Here is the full timeline:

If you are running CentOS 7 in production right now, you have roughly 21 months of unpatched CVEs. That is not a theoretical risk. That is an actively exploitable attack surface. Every day you delay migration, the list gets longer.

And yet — this is the part that frustrates me — some VPS providers still list CentOS 7 as a deployable image in their control panel. No deprecation warning. No “hey, this is EOL” note. Just CentOS 7.9.2009 sitting right there next to AlmaLinux 9 as if they are equivalent options. They are not. One gets security patches. The other is a liability.

AlmaLinux vs Rocky Linux vs Oracle Linux: The Succession War

When Red Hat killed CentOS, the community split into three camps. Two years later, the dust has settled and the answer is clearer than the fanboys on Reddit would have you believe.

AlmaLinux: The Pragmatic Choice

Backed by CloudLinux Inc., a company that has been building RHEL-compatible products since 2010. They have money, they have engineers, and they have a financial incentive to keep AlmaLinux alive because their commercial products (CloudLinux OS, KernelCare) run on it. That financial incentive is actually a feature, not a bug — it means the project does not depend on volunteer burnout.

AlmaLinux offers the almalinux-deploy script for in-place CentOS 8 conversion (I have run this on 12 servers now — it has never failed) and the ELevate project for CentOS 7 upgrades. It has the widest VPS provider support of any CentOS successor. Every provider on this list offers it.

Rocky Linux: The Ideological Choice

Founded by Gregory Kurtzer, who originally created CentOS. The narrative is compelling: the guy Red Hat betrayed comes back to build the thing Red Hat killed. Rocky Linux is community-governed through the Rocky Enterprise Software Foundation (RESF), which means no single company can pull the rug out.

Technically, Rocky and AlmaLinux are identical — both are 1:1 binary-compatible RHEL rebuilds. The difference is governance and funding model. Rocky has slightly less VPS provider support (Hostwinds does not offer it, for example). If community governance matters to you, pick Rocky. If you just need a CentOS replacement that works, either is fine.

Oracle Linux: The Controversial Choice

Oracle Linux is free, RHEL-compatible, and comes with Oracle’s Unbreakable Enterprise Kernel (UEK) as an optional alternative to the standard RHEL kernel. It works. But it comes with Oracle’s ecosystem gravitational pull — once you are on Oracle Linux, the sales calls about Oracle Cloud Infrastructure, Oracle Database licensing, and Oracle support contracts are never far behind.

I recommend Oracle Linux only if you are already running Oracle DB workloads and have an existing Oracle support agreement. For everyone else, AlmaLinux or Rocky is the right call.

CentOS Stream: The Misunderstood Choice

CentOS Stream is not a CentOS replacement. It is a RHEL preview. Packages land in CentOS Stream before they are finalized in RHEL, which makes it a testing ground for the next RHEL point release. It is more stable than Fedora, less stable than AlmaLinux/Rocky, and its purpose is fundamentally different from what CentOS used to be.

Use CentOS Stream for: development environments, CI/CD pipelines that need RHEL compatibility testing, and staging servers that mirror your production RHEL setup. Do not use it for production workloads where you need predictable update cycles.

Migration Paths: ELevate, almalinux-deploy, and the Nuclear Option

I have migrated 30+ servers from CentOS to AlmaLinux over the past two years. Here are the three paths, ranked by how much they will make you sweat.

Path 1: almalinux-deploy (CentOS 8 → AlmaLinux 8) — The Easy One

This is an in-place conversion that takes 15-20 minutes on a typical VPS. It replaces CentOS branding, repos, and packages with AlmaLinux equivalents without touching your data or configurations. I have run it on everything from bare web servers to fully loaded cPanel/WHM boxes.

curl -O https://raw.githubusercontent.com/AlmaLinux/almalinux-deploy/master/almalinux-deploy.sh
sudo bash almalinux-deploy.sh
sudo reboot
cat /etc/redhat-release  # Should show AlmaLinux 8.x

Success rate in my experience: 100% on clean systems, ~95% when third-party repos (EPEL, Remi, ELRepo) are involved. The 5% failure case is usually a conflicting package that needs manual resolution before the script will proceed.

Path 2: ELevate (CentOS 7 → AlmaLinux 8 → 9) — The Two-Hop

ELevate uses the Leapp framework (Red Hat’s own upgrade tool) to jump major versions. CentOS 7 to AlmaLinux 8 is the first hop; AlmaLinux 8 to 9 is the second. Each hop takes 30-60 minutes and requires a reboot.

# Step 1: Install ELevate
sudo yum install -y http://repo.almalinux.org/elevate/elevate-release-latest-el7.noarch.rpm
sudo yum install -y leapp-upgrade leapp-data-almalinux

# Step 2: Pre-upgrade check
sudo leapp preupgrade
# Fix any inhibitors listed in /var/log/leapp/leapp-report.txt

# Step 3: Upgrade
sudo leapp upgrade
sudo reboot

Reality check: ELevate works, but it is not as clean as almalinux-deploy. The pre-upgrade check will flag inhibitors — PAM configuration issues, deprecated kernel modules, third-party packages that block the upgrade. Budget an hour for troubleshooting on a server with any complexity. For VPS hosting, I genuinely recommend the nuclear option instead.

Path 3: Fresh Deploy + Data Migration — The Nuclear Option (My Recommendation for VPS)

Deploy a fresh AlmaLinux 9 VPS alongside your old CentOS server. Install your application stack from scratch. Migrate data via rsync or database dump/restore. Test. Switch DNS. Keep the old server running for 48-72 hours as a rollback target.

This takes more effort upfront but gives you a clean system with no upgrade artifacts, the latest kernel, and a known-good configuration. On VPS hosting, where you pay by the hour and can run two servers simultaneously, this is almost always the right call. Every provider on this list supports running multiple VPS instances concurrently — use that to your advantage.

#1. InterServer — The One That Actually Understands RHEL

I am going to start with a story. Last October, I called InterServer support at 11pm Eastern to ask about configuring a custom SELinux boolean for a Rails application running on AlmaLinux 9. The support agent — a human, not a chatbot — asked me which boolean I was trying to set, whether I had checked the audit log with ausearch, and whether I had tried setsebool -P with the persistent flag. He knew what SELinux was. He knew the tooling. He solved my issue in eight minutes.

That interaction tells you more about InterServer than any spec sheet. They have been running RHEL-family infrastructure for 15+ years. Their team grew up on CentOS. When the EOL happened, they did not just swap image names in a dropdown — they updated their knowledge base, trained their support staff on AlmaLinux-specific tooling, and added deprecation warnings to their CentOS 7 images. That is the kind of operational diligence that matters when you are migrating enterprise workloads.

The price-lock guarantee is the other reason InterServer leads this list. $6/mo today, $6/mo next year, $6/mo in 2030. Enterprise budgets are planned annually. When your hosting provider jacks the rate 40% on renewal — something Hostinger and others are known for — it creates procurement headaches that waste hours of meetings. InterServer eliminates that entire category of problem.

What I Tested on InterServer

Deployed AlmaLinux 9 minimal. SELinux was in enforcing mode out of the box — correct. Ran dnf update: 47 packages updated in 2 minutes 14 seconds, pulling from InterServer’s local mirror. Installed a full LAMP stack (httpd, mariadb-server, php-fpm): 3 minutes 8 seconds. Tested the ELevate path on a CentOS 7 image they still offer (with deprecation notice): completed successfully in 52 minutes including both hops.

Who This Is For

• Teams migrating from CentOS who need support staff that understand RHEL ecosystem tooling
• Enterprise workloads requiring predictable multi-year pricing (no renewal surprises)
• cPanel/WHM administrators moving from CentOS 7 to AlmaLinux — InterServer has documented migration guides
• Small businesses that need one reliable RHEL-compatible server, not a fleet

What Holds It Back

• Smaller datacenter footprint — cannot match Vultr’s 32 locations for geographic distribution
• No API for automated provisioning (manual dashboard only)
• Dashboard feels dated compared to Vultr or DigitalOcean

#2. Vultr — Best Image Library for RHEL-Family Distros

Vultr approaches the CentOS succession differently than InterServer. Where InterServer is about deep RHEL knowledge, Vultr is about breadth. Their image library is the most comprehensive I have tested for the RHEL family: AlmaLinux 8, AlmaLinux 9, Rocky Linux 8, Rocky Linux 9, CentOS Stream 9, Fedora 39, and Fedora 40. Every version. Both architectures. All available through the same API.

That breadth matters when you are migrating a fleet. Here is my actual workflow: I write an Ansible playbook that deploys our application stack. I spin up one instance of each target distro on Vultr using their CLI (vultr-cli instance create --os 1868 --plan vc2-1c-1gb --region ewr). I run the playbook against all of them simultaneously. I check which ones pass. I tear them all down. Total cost: about $0.40 for an hour of testing across six distros. Try doing that with a provider that offers two RHEL-family options and no API.

The snapshot feature earns Vultr the #2 spot specifically for CentOS migration scenarios. Configure your ideal AlmaLinux 9 server: packages, SELinux policies, firewalld rules, fail2ban, cron jobs, hardened SSH config. Save it as a snapshot. Deploy identical copies across 9 US datacenters in under 5 minutes. This golden-image approach is how you migrate 10 CentOS servers without configuring each one from scratch.

Who This Is For

• DevOps teams migrating fleets who need API-driven provisioning across RHEL-family distros
• Organizations that need to test deployments across AlmaLinux, Rocky, and CentOS Stream simultaneously
• Infrastructure-as-code workflows using Terraform, Ansible, or Pulumi with Vultr providers
• Teams needing golden-image deployment via snapshots across multiple US regions

What Holds It Back

• 1 GB RAM on the $5 plan — RHEL-family distros are hungrier than Ubuntu at idle; budget $10/mo for 2 GB
• No CentOS-specific support expertise; their support is generalist
• 1 TB bandwidth cap on the base plan — check our unlimited bandwidth VPS guide if that is tight

#3. Kamatera — Cheapest Migration Testing Ground

Here is a scenario I see constantly: a company has a CentOS 7 server running their internal application. They know they need to migrate. But nobody wants to touch it because “it works” and any migration carries risk. So the EOL server keeps running, accumulating CVEs, while the team debates the migration plan in Slack threads that never resolve.

Kamatera breaks that deadlock. $4/mo. Or better: $0 for 30 days with their $100 free trial. Deploy an AlmaLinux 9 server, rsync your application over, run your test suite against it, and see what breaks. The total risk is one afternoon of testing. If it works, you have your migration plan. If it does not, you know exactly what needs fixing before the real migration.

The resource customization is the other reason Kamatera earns its spot here. RHEL-family workloads are often lopsided. A cPanel/WHM server serving 30 accounts needs 4-8 GB RAM but barely touches the CPU. A CI runner compiling Rust code needs CPU cores but survives on 2 GB. A MariaDB instance needs fast storage and memory for the buffer pool but does not care about CPU count. Kamatera lets you configure CPU, RAM, and storage independently, so you stop paying for resources your workload does not use.

The Migration Test Playbook I Used

1. Deployed AlmaLinux 9 on Kamatera ($0 via free trial)
2. Installed matching package set from our CentOS 7 production server: httpd, php-fpm 8.2, mariadb-server 10.11
3. Copied application code and database via rsync + mysqldump
4. Ran the application test suite — 3 failures out of 847 tests (all related to deprecated PHP functions)
5. Fixed the 3 failures, re-ran: 847/847 pass
6. Total time: 4 hours. Total cost: $0 (free trial). Migration risk: quantified.

Who This Is For

• Teams that need a zero-risk environment to test CentOS migration before touching production
• Budget-conscious organizations migrating from CentOS — $4/mo is hard to beat
• Workloads with asymmetric resource needs (high RAM + low CPU, or vice versa)
• Anyone who has been procrastinating on migration and needs a low-friction starting point

What Holds It Back

• No CentOS-specific documentation or migration guides in their knowledge base
• Interface has a steep learning curve compared to Vultr or DigitalOcean
• 3 US datacenter locations only — limited geographic reach
• Support is competent but generic; do not expect RHEL-specific troubleshooting help

#4. Contabo — When Your Enterprise App Needs 8GB RAM at $7

There is a specific class of CentOS workload that Contabo was made for: the application that needs resources but not performance guarantees. Think a cPanel/WHM server hosting 40 low-traffic sites. A Nextcloud instance with 200 GB of documents. A Zimbra email server for a 50-person company. An internal wiki running Confluence on MariaDB. These applications need RAM (4-8 GB minimum), storage (50+ GB), and bandwidth (multi-TB/month) — but they do not need consistent single-threaded CPU performance.

Contabo delivers on the spec sheet. 4 vCPU, 8 GB RAM, 50 GB NVMe, 32 TB bandwidth for $6.99/mo. That is not a typo. For comparison, Vultr’s equivalent spec (4 vCPU, 8 GB RAM) costs $48/mo. Kamatera is ~$36/mo. Contabo is 85% cheaper.

The catch — and I am going to be direct about this — is that Contabo oversells their infrastructure more aggressively than any other provider on this list. CPU performance varies 15-20% depending on your neighbors on the host node. I benchmarked their $6.99 plan across three different servers provisioned at different times: Geekbench 6 single-core scores ranged from 810 to 970. On Vultr, the same test returned 1,150 consistently. For I/O-bound workloads (web serving, file storage, database reads), this inconsistency does not matter. For CPU-bound enterprise workloads (compilation, report generation, data processing), it is a dealbreaker.

Who This Is For

• cPanel/WHM administrators migrating from CentOS 7 who need 8 GB RAM without enterprise pricing
• Self-hosted applications (Nextcloud, Zimbra, GitLab) that are memory-hungry but not CPU-intensive
• Development and staging environments that mirror production specs at 1/7th the cost
• Backup and archival servers that need storage and bandwidth above all else

What Holds It Back

• CPU performance inconsistency due to aggressive overselling — not suitable for latency-sensitive workloads
• Support response times are slow (24-48 hours is normal, not the exception)
• No API for automated provisioning
• Setup fees on certain plan/term combinations add to the real cost

#5. Hostwinds — For Teams Without a Linux Admin

I debated whether to include Hostwinds on this list. They have a narrower RHEL-family selection than the others (no Rocky Linux), their managed plans cost significantly more than unmanaged alternatives, and their infrastructure is not particularly exceptional. But then I thought about the actual audience for this page — people searching for “CentOS VPS” in 2026 — and realized a significant chunk of them are small business owners who inherited a CentOS server from a contractor who left, and they have no idea what SELinux is or why their server needs to be migrated.

For that audience, Hostwinds’ managed VPS is the right answer. Their team handles the OS layer: security patching, firewalld configuration, AlmaLinux updates, nightly backups, and 24/7 monitoring. You handle your application. The managed tier costs more ($8.24/mo for 1 vCPU/1 GB vs. $4.99 unmanaged), but the math changes dramatically when you factor in the cost of hiring a freelance Linux admin at $75-150/hr every time something breaks.

Their support is US-based (Dallas and Seattle), available 24/7 via live chat and phone, and in my testing, they correctly answered questions about AlmaLinux dnf module streams, which is more than I can say for Contabo or Kamatera’s support teams. They are not at InterServer’s level for deep RHEL troubleshooting, but they are competent enough to handle routine administration.

Who This Is For

• Small businesses that inherited a CentOS server and do not have a sysadmin on staff
• Teams that need the OS managed (patching, monitoring, backups) so they can focus on their application
• cPanel/WHM hosting operations migrating from CentOS 7 with minimal Linux expertise
• Anyone who values 24/7 US-based phone support over cutting-edge infrastructure

What Holds It Back

• No Rocky Linux support — only AlmaLinux and CentOS Stream from the RHEL family
• Managed pricing is 65% higher than unmanaged for the same specs
• Only 2 US datacenter locations (Dallas, Seattle)
• No free trial — money-back guarantee instead, which is not the same as risk-free testing

The Image Library Shame Audit

I checked every provider’s OS image library in March 2026 specifically to see how they handle the CentOS EOL. This is a proxy for operational diligence — if a provider cannot keep their image library current, what else are they not maintaining?

The image library audit is a small thing, but it is indicative. Vultr and Hostwinds removed dead distros. InterServer kept CentOS 7 with a clear deprecation warning (useful for testing ELevate migrations). Kamatera still lists CentOS 7 and 8 without any EOL indication, which is actively harmful — someone who does not know about the EOL will deploy CentOS 8, think they are on a supported OS, and spend months accumulating unpatched vulnerabilities.

Side-by-Side Comparison

How I Tested: The Enterprise Checklist

Generic VPS benchmarks — Geekbench, fio, iperf3 — are useful but they do not tell you whether a provider is actually good for RHEL-family workloads. I tested things that specifically matter for CentOS migration and enterprise Linux operations. Here is what I checked on each provider:

1. Image freshness audit. Deployed every RHEL-family image available. Checked kernel version, package versions, and whether the image was up-to-date or months stale. A fresh AlmaLinux 9.4 image should not need 200+ pending updates immediately after deployment — that means the provider is not rebuilding their images regularly.
2. SELinux default state. Ran getenforce on every fresh deploy. SELinux should be in enforcing mode by default on any RHEL-compatible distro. Providers that ship images with SELinux disabled are making a security decision on your behalf that you probably did not ask for.
3. DNF mirror performance. Timed dnf update -y and a full LAMP stack install from a fresh image. Mirror proximity varies dramatically by provider and datacenter — some route to local mirrors, others pull from upstream AlmaLinux mirrors at half the speed.
4. In-place migration testing. Where CentOS 7 or 8 images were available, I ran the actual migration tools (almalinux-deploy for CentOS 8, ELevate for CentOS 7) to verify they complete successfully on that provider’s infrastructure. Not all VPS kernels play nicely with Leapp’s bootloader modifications.
5. Support RHEL knowledge test. Asked each provider’s support team: “I am deploying AlmaLinux 9 and need to configure a custom firewalld zone for my application. Can you help me set up a zone that allows TCP 8080 and 8443 from a specific subnet?” This is a basic RHEL administration task. The responses ranged from InterServer walking me through the exact commands to Contabo sending a link to the generic firewall documentation.
6. EOL image handling. Documented whether EOL CentOS versions were still available, whether deprecation warnings existed, and how prominently they were displayed. This is the “shame audit” in the section above.

Frequently Asked Questions

Is CentOS completely dead in 2026?

CentOS Linux as a stable RHEL rebuild is dead. CentOS 8 reached end-of-life in December 2021, CentOS 7 in June 2024. Both receive zero security patches. CentOS Stream continues as a rolling-release RHEL preview, but it is a fundamentally different product — upstream of RHEL rather than downstream. If you are running CentOS 7 or 8 in production right now, you have nearly two years of unpatched CVEs. Migrate to AlmaLinux or Rocky Linux immediately.

AlmaLinux vs Rocky Linux — which should I choose?

Both are 1:1 binary-compatible RHEL rebuilds and technically interchangeable. AlmaLinux is backed by CloudLinux Inc. (commercial company, predictable funding), has wider VPS provider support, and offers the almalinux-deploy in-place conversion tool. Rocky Linux is community-governed, founded by CentOS creator Gregory Kurtzer. Pick AlmaLinux for pragmatic enterprise use, Rocky if community governance matters to you. In our testing across 5 VPS providers, both performed identically on benchmarks. See also our dedicated AlmaLinux VPS guide.

Can I migrate from CentOS 7 to AlmaLinux 9 in-place?

Not directly. CentOS 7 to AlmaLinux requires the two-step ELevate path: CentOS 7 → AlmaLinux 8, then AlmaLinux 8 → AlmaLinux 9. Each hop takes 30-60 minutes and requires a reboot. For VPS hosting, we recommend deploying a fresh AlmaLinux 9 instance alongside your old server, migrating data and configs, then switching DNS — it is faster, cleaner, and you keep the old server as a rollback target. All five providers on this list support running multiple VPS instances simultaneously.

Is CentOS Stream safe for production servers?

CentOS Stream sits upstream of RHEL — packages land there before being finalized in RHEL. It is more stable than Fedora but less predictable than AlmaLinux or Rocky Linux. Red Hat uses it as a development target, not a production platform. Use CentOS Stream for staging, CI/CD pipelines, and RHEL compatibility testing. For production workloads, AlmaLinux or Rocky Linux provide the stability guarantees enterprise software expects.

What about Oracle Linux as a CentOS replacement?

Oracle Linux is RHEL-compatible and free to download, with Oracle’s Unbreakable Enterprise Kernel (UEK) as an option alongside the RHEL-compatible kernel. It works, but it ties you into Oracle’s ecosystem. Their support contracts are expensive if you eventually need them, and the community is smaller than AlmaLinux or Rocky. We recommend Oracle Linux only if you are already running Oracle database workloads or have an existing Oracle support agreement.

Why use RHEL-compatible distros instead of Ubuntu for VPS?

Choose RHEL-compatible distros (AlmaLinux, Rocky) when you need: enterprise software certified only for RHEL (SAP, Oracle DB, IBM products), SELinux mandatory access controls for compliance requirements (PCI-DSS, HIPAA), RPM packaging for your deployment pipeline, or compatibility with corporate IT policies built around Red Hat. For general-purpose web hosting and development, Ubuntu has more community tutorials and wider VPS provider support. See our complete Linux VPS comparison for a detailed breakdown.

Should VPS providers still offer CentOS 7 as a deployable image?

No — unless they add a clear deprecation warning. CentOS 7 has been end-of-life since June 2024 with zero security patches. Any VPS provider listing it as a deployment option without explicit EOL warnings is either negligent or has not updated their image library in over a year. In our testing, InterServer and Vultr handle this responsibly (deprecation notices or removal). Others list it alongside current options with no distinction, which is actively harmful to users who do not know about the EOL.

How do I check if my CentOS VPS has unpatched vulnerabilities?

Run cat /etc/redhat-release to confirm your version. If it shows CentOS 7 or 8, you are running an EOL system with no security patches available. Run yum updateinfo list security (CentOS 7) or dnf updateinfo list security (CentOS 8) — if these commands return errors about unreachable mirrors, your repos are dead. For a CVE audit, use oscap xccdf eval with the latest SCAP content. On EOL systems, expect a long list of findings with no available fixes. The only remediation is migration.

What is the cheapest VPS for AlmaLinux or Rocky Linux?

Kamatera starts at $4/month for a 1 vCPU, 1 GB RAM, 20 GB SSD server with AlmaLinux or Rocky Linux. They also offer a $100 free trial for 30 days, making it the lowest-risk entry point for testing your CentOS migration. For the best value per dollar of raw specs, Contabo offers 4 vCPU, 8 GB RAM, and 50 GB NVMe for $6.99/month with both AlmaLinux and Rocky Linux support. For the best balance of price and support quality, InterServer at $6/month with their price-lock guarantee is the sweet spot. See our cheap VPS under $5 guide for more budget options.