Re: [PVE-User] API redundancy

[Stefan Radman via pve-user <pve-user@lists.proxmox.com>]

[-- Attachment #1: Type: message/rfc822, Size: 7986 bytes --]

From: Stefan Radman <stefan.radman@me.com>
To: Proxmox VE user list <pve-user@lists.proxmox.com>
Subject: Re: [PVE-User] API redundancy
Date: Fri, 14 Mar 2025 20:36:06 +0100
Message-ID: <C680BA8B-9899-42AC-96BD-1D206BEC8021@me.com>

Hi Cyrus

Here is the simple keepalived configuration we're using.

It serves us very well for redundancy of the web GUI.

All you need is “apt install keepalived” on every cluster node and the configuration in /etc/keepalived/keepalived.conf.

Just mind the different priorities - gory details in the keepalived.conf(1) man page.

The browser title page will show you which node you are actually connected to.

Stefan

root@pve01:~# host pve && echo && for node in pve01 pve02 pve03; do host $node && ssh $node cat /etc/keepalived/keepalived.conf && echo; done
pve.example.com has address 10.10.10.10

pve01.example.com has address 10.10.10.1
vrrp_instance pve {
    interface vmbr1
    virtual_router_id 10
    priority 255
    virtual_ipaddress {
        10.10.10.10/24
    }
}

pve02.example.com has address 10.10.10.2
vrrp_instance pve {
    interface vmbr1
    virtual_router_id 10
    priority 254
    virtual_ipaddress {
        10.10.10.10/24
    }
}

pve03.example.com has address 10.10.10.3
vrrp_instance pve {
    interface vmbr1
    virtual_router_id 10
    priority 253
    virtual_ipaddress {
        10.10.10.10/24
    }
}


> On Mar 14, 2025, at 19:35, Roberto Alvarado <ralvarado@anycast.cl> wrote:
> 
> No, in proxmox this kind of sceneraio isnt resolved, other way (without a reverse proxy) can be done with keepalived (vrrp), and use a "floating IP"
> 
> Saludos
> Roberto
> 
> On 14-03-25 15:19, Cyrus wrote:
>> Hello!,
>> 
>> Yes, I can do it outside with a reverse proxy.
>> 
>> The message I'm rescuing: there's no transparent HA included out of the box
>> to cover that scenario
>> 
>> Regards.
>> 
>> On Fri, Mar 14, 2025, 14:59 Roberto Alvarado <ralvarado@anycast.cl> wrote:
>> 
>>> You can do this with nginx or haproxy, using each pve server as backend.
>>> 
>>> 
>>> Saludos
>>> Roberto
>>> 
>>> On 14-03-25 13:59, Cyrus wrote:
>>>> Hello!,
>>>> 
>>>> Understanding there's not cluster IP (failover), how do you provide
>>>> redundancy for API consumers when one node is not accesible?
>>>> 
>>>> (User configured with API endpoint https://pve01:8006, pve01 going down,
>>>> pve02 accesible)
>>>> 
>>>> Regards.
>>>> _______________________________________________
>>>> pve-user mailing list
>>>> pve-user@lists.proxmox.com
>>>> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user
>>>> 
>>> _______________________________________________
>>> pve-user mailing list
>>> pve-user@lists.proxmox.com
>>> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user
>>> 
>>> 
>> _______________________________________________
>> pve-user mailing list
>> pve-user@lists.proxmox.com
>> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user
>> 
> 
> _______________________________________________
> pve-user mailing list
> pve-user@lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user
> 


[-- Attachment #2: Type: text/plain, Size: 157 bytes --]

_______________________________________________
pve-user mailing list
pve-user@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user