17.2. pveproxy - Proxmox VE API Proxy Daemon

This daemon exposes the whole Proxmox VE API on TCP port 8006 using HTTPS. It runs as user www-data and has very limited permissions. Operation requiring more permissions are forwarded to the local pvedaemon.

Requests targeted for other nodes are automatically forwarded to those nodes. This means that you can manage your whole cluster by connecting to a single Proxmox VE node.

It is possible to configure “apache2”-like access control lists. Values are read from file /etc/default/pveproxy. For example:

ALLOW_FROM="10.0.0.1-10.0.0.5,192.168.0.0/22"
DENY_FROM="all"
POLICY="allow"

IP addresses can be specified using any syntax understood by Net::IP. The name all is an alias for 0/0 and ::/0 (meaning all IPv4 and IPv6 addresses).

The default policy is allow.

Match POLICY=deny POLICY=allow

Match Allow only

allow

allow

Match Deny only

deny

deny

No match

deny

allow

Match Both Allow & Deny

deny

allow

By default the pveproxy and spiceproxy daemons listen on the wildcard address and accept connections from both IPv4 and IPv6 clients.

By setting LISTEN_IP in /etc/default/pveproxy you can control to which IP address the pveproxy and spiceproxy daemons bind. The IP-address needs to be configured on the system.

Setting the sysctl net.ipv6.bindv6only to the non-default 1 will cause the daemons to only accept connection from IPv6 clients, while usually also causing lots of other issues. If you set this configuration we recommend to either remove the sysctl setting, or set the LISTEN_IP to 0.0.0.0 (which will only allow IPv4 clients).

LISTEN_IP can be used to only to restricting the socket to an internal interface and thus have less exposure to the public internet, for example:

LISTEN_IP="192.0.2.1"

Similarly, you can also set an IPv6 address:

LISTEN_IP="2001:db8:85a3::1"

Note that if you want to specify a link-local IPv6 address, you need to provide the interface name itself. For example:

LISTEN_IP="fe80::c463:8cff:feb9:6a4e%vmbr0"

To apply the change you need to either reboot your node or fully restart the pveproxy and spiceproxy service:

systemctl restart pveproxy.service spiceproxy.service

You can define the cipher list in /etc/default/pveproxy, for example

CIPHERS="ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA256"

Above is the default. See the ciphers(1) man page from the openssl package for a list of all available options.

Additionally, you can set the client to choose the cipher used in /etc/default/pveproxy (default is the first cipher in the list available to both client and pveproxy):

HONOR_CIPHER_ORDER=0

You can define the used Diffie-Hellman parameters in /etc/default/pveproxy by setting DHPARAMS to the path of a file containing DH parameters in PEM format, for example

DHPARAMS="/path/to/dhparams.pem"

If this option is not set, the built-in skip2048 parameters will be used.

You can change the certificate used to an external one or to one obtained via ACME.

pveproxy uses /etc/pve/local/pveproxy-ssl.pem and /etc/pve/local/pveproxy-ssl.key, if present, and falls back to /etc/pve/local/pve-ssl.pem and /etc/pve/local/pve-ssl.key. The private key may not use a passphrase.

See the Host System Administration chapter of the documentation for details.

By default pveproxy uses gzip HTTP-level compression for compressible content, if the client supports it. This can disabled in /etc/default/pveproxy

COMPRESSION=0