From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from firstgate.proxmox.com (firstgate.proxmox.com [212.224.123.68]) by lore.proxmox.com (Postfix) with ESMTPS id 2797F1FF13C for ; Thu, 11 Jun 2026 13:35:47 +0200 (CEST) Received: from firstgate.proxmox.com (localhost [127.0.0.1]) by firstgate.proxmox.com (Proxmox) with ESMTP id C19FAFA; Thu, 11 Jun 2026 13:35:32 +0200 (CEST) Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=UTF-8 Date: Thu, 11 Jun 2026 13:34:55 +0200 Message-Id: From: "Max R. Carrara" Subject: Re: Question about Proxmox VE stretched cluster between two datacenters To: =?utf-8?q?Iv=C3=A1n_M=C3=A9ndez_Cair=C3=B3s?= , "pve-user@lists.proxmox.com" X-Mailer: aerc 0.18.2-0-ge037c095a049 References: In-Reply-To: X-Bm-Milter-Handled: 55990f41-d878-4baa-be0a-ee34c49e34d2 X-Bm-Transport-Timestamp: 1781177647808 X-SPAM-LEVEL: Spam detection results: 0 AWL 0.081 Adjusted score from AWL reputation of From: address BAYES_00 -1.9 Bayes spam probability is 0 to 1% DMARC_MISSING 0.1 Missing DMARC policy KAM_DMARC_STATUS 0.01 Test Rule for DKIM or SPF Failure with Strict Alignment SPF_HELO_NONE 0.001 SPF: HELO does not publish an SPF Record SPF_PASS -0.001 SPF: sender matches SPF record URIBL_BLOCKED 0.001 ADMINISTRATOR NOTICE: The query to URIBL was blocked. See http://wiki.apache.org/spamassassin/DnsBlocklists#dnsbl-block for more information. [proxmox.com] Message-ID-Hash: ZW5UQZYGSS2EOXNXBFNWHJ7VBKOQK54O X-Message-ID-Hash: ZW5UQZYGSS2EOXNXBFNWHJ7VBKOQK54O X-MailFrom: m.carrara@proxmox.com X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; loop; banned-address; emergency; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header X-Mailman-Version: 3.3.10 Precedence: list List-Id: Proxmox VE user list List-Help: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: On Thu Jun 11, 2026 at 12:17 PM CEST, Iv=C3=A1n M=C3=A9ndez Cair=C3=B3s wro= te: > Hello everyone, > > We would like to know if it is possible to implement a stretched cluster = with Proxmox VE in the following scenario: > > * Two datacenters (Production and Backup/DR) > * Same infrastructure on both sites > * Dedicated virtualization servers on both datacenters > * Dedicated storage arrays on both datacenters > * High Availability between both sites > * Automatic failover between datacenters > > The idea is that, if the main datacenter goes down completely, the backup= datacenter automatically continues providing the services without manual i= ntervention. > > Is this possible with Proxmox VE? > If so, what would be the recommended architecture or best practices for t= his kind of deployment? > > Thank you very much. > > Best regards, > > Iv=C3=A1n Hi Iv=C3=A1n! Yes, that is indeed possible with Proxmox VE + Ceph. PVE tightly integrates Ceph, resulting in a fully hyper-converged infrastructure (HCI). The "stretch cluster" is what you want, and we do in fact have a guide on that: https://pve.proxmox.com/wiki/Stretch_Cluster Note that in a HCI cluster storage and compute aren't separated as with "traditional" setups, so you generally wouldn't have dedicated virtualization servers and dedicated storage arrays. You probably know this already, but I wanted to mention it nevertheless. More on HCI here: https://pve.proxmox.com/wiki/Hyper-converged_Infrastructure So, to address your points individually: > * Two datacenters (Production and Backup/DR) Done with Ceph with appropriate CRUSH rules, but you do need a third location for a tie-breaker node (see the stretch cluster guide above). What's also possible is having three datacenters instead, but I assume that's not within your budget. ;) > * Same infrastructure on both sites Yeah, that works excellently with the stretch cluster setup. There are no restrictions when it comes to hardware, network topology, etc., so how *exactly* you define your infrastructure is up to you. > * Dedicated virtualization servers on both datacenters > * Dedicated storage arrays on both datacenters As mentioned earlier, storage and compute aren't separated anymore; all nodes work as compute and storage nodes. That means no "storage boxes" etc. This might seem a bit strange / unfamiliar to some, but the main benefit is that you can scale up super easily. Ceph performs better the more nodes and OSDs you add and, with appropriate configuration (CRUSH rules etc.), ensures that your data is replicated across both datacenters. The exact configuration is up to you, but the stretch cluster guide does show how you can have two replicas per location -- meaning that each datacenter contains two copies of your data. > * High Availability between both sites > * Automatic failover between datacenters Done with our HA Manager: https://pve.proxmox.com/wiki/High_Availability So yes, if one datacenter were to go down for some reason, the other one would just continue chugging along, no manual intervention required. That should answer all of your questions, I think. Hope that helps! Best regards, Max