corosync-qnetd - Man Page

QNet daemon

Synopsis

corosync-qnetd [-46dfhv] [-l listen_addr] [-p listen_port] [-s tls] [-c client_cert_required] [-m max_clients] [-S option=value[,option2=value2,...]]"

corosync-qnetd is a daemon running outside of the cluster with the purpose of providing a vote to the corosync-qdevice model net. It's designed to support multiple clusters and be almost configuration and state free. New clusters are handled dynamically and no configuration file exists. It's also able to run as non-root user - which is recommended. Connection between the corosync-qdevice model net client can be optionally configured with TLS client certificate checking. The communication protocol between server and client is designed to be very simple and allow backwards compatibility.

Options

-4: and its counterpart -6 are used to force IPv4 or IPv6 communication. The default is to listen on both address families.
-d: Turn on debug logging. By default the messages sent to syslog are purely operational, this option sends additional debug messages. For bumping syslog messages priority to info, use the -d parameter twice.
-f: Do not daemonize, run in the foreground.
-h: Show short help text
-v: Show version and supported communication protocol messages/options.
-l: IP address to listen on. By default the daemon listens on all addresses (wildcard).
-p: TCP port to listen on. Default port is 5403.
-s: Determines if TLS should be used and can be one of on/off/required (the default is on ). on means TLS is enabled but the client is not required to start TLS, off means TLS is completely disabled, and required means TLS is required. on and required require the NSS database to be properly initialized by running the corosync-qnetd-certutil command.
-c: can be set to on/off. This option only makes sense if TLS is enabled. When -c is on a client is required to send its client certificate (default).
-m: Maximum simultaneous clients. The default is 0 which means no limit.
-S: Set advanced settings described in its own section below. This option shouldn't be generally used because most of the options are not safe to change.

Unprivileged User Configuration

It's generally recommended to run corosync-qnetd as a non root user. If you get a package from a distribution its highly possible that the packager has done all the hard work for you. If the installation is performed from source code, a few steps have to be taken.

First it's necessary to create an unprivileged user/group. The following commands can be used (executed as root):

# groupadd -r coroqnetd
# useradd -r -g coroqnetd -d / -s /sbin/nologin -c "User for corosync-qnetd" coroqnetd

The next step is to set the correct owner and group on /etc/corosync/qnetd and /var/run/corosync-qnetd directories.

# chown -R coroqnetd:coroqnetd /etc/corosync/qnetd /var/run/corosync-qnetd

Some systems have the /var/run directory on a tmpfs file system which gets discarded after a reboot. The solution is to use an initscript or systemd unit, because both of them takes care of the /var/run/corosync-qnetd creation and sets the correct owner and permissions.

The last step is to make sure corosync-qnetd is really executed as an unprivileged user. For initscript systems it's enough to set the line COROSYNC_QNETD_RUNAS in /etc/(sysconfig|default)/corosync-qnetd file. If the file is not already installed then use the one provided in the corosync source code (init/corosync-qnetd.sysconfig.example). For systemd, overwrite/copy the corosync-qnetd.service unit file and uncomment/change the "User=" directive.

TLS Configuration

For TLS to work its necessary to create the NSS database. If pcs is used then the following steps are not needed because pcs does them automatically.

corosync-qnetd-certutil is the tool to perform required actions. Just run:

# corosync-qnetd-certutil -i

If TLS is not required then simply edit /etc/(sysconfig|default)/corosync-qnetd or systemd unit file and add the parameter -s off in the proper place.

Depending on configuration of NSS (stored in nss.config file usually in /etc/crypto-policies/back-ends/ directory) disabled ciphers or too short keys may be rejected. Proper solution is to regenerate NSS databases for both corosync-qnetd and corosync-qdevice daemons. As a quick workaround it's also possible to set environment variable NSS_IGNORE_SYSTEM_POLICY=1 before running corosync-qnetd daemon.

When NSS is updated it may also be needed to upgrade database into new format. There is no consensus on recommended way, but following command seems to work just fine (if qnetd sysconfdir is set to /etc)

# certutil -N -d /etc/corosync/qnetd/nssdb -f /etc/corosync/qnetd/nssdb/pwdfile.txt

Advanced Settings

Set by the -S option. The default value is shown in parentheses.

listen_backlog: Parameter passed to the listen syscall on the network socket. (10)
max_client_send_buffers: Maximum number of send buffers for one client. (32)
max_client_send_size: Maximum size of one send buffer (message) to be sent to a client. (32768)
max_client_receive_size: Maximum size of the receive buffer for a client message (maximum allowed message size received by client). (32768)
nss_db_dir: NSS database directory. (/etc/corosync/qnetd/nssdb)
cert_nickname: NSS nickname of qnetd server certificate. (QNetd Cert)
heartbeat_interval_min: Minimum heartbeat timeout accepted by server in ms. (1000)
heartbeat_interval_max: Maximum heartbeat timeout accepted by server in ms. (120000)
dpd_enabled: Dead peer detection enabled. (on)
dpd_interval_coefficient: Value is multiplied with heartbeat interval sent by qdevice client and used as a timeout for dead peer detection. (1.5)
lock_file: Lock file location. (/var/run/corosync-qnetd/corosync-qnetd.pid)
local_socket_file: Internal IPC socket file location. (/var/run/corosync-qnetd/corosync-qnetd.sock)
local_socket_backlog: Parameter passed to listen syscall on the local socket. (10)
ipc_max_clients: Maximum allowed simultaneous IPC clients. (10)
ipc_max_receive_size: Maximum size of a message received by IPC client. (4096)
ipc_max_send_size: Maximum size of a message sent to an IPC client. (10485760)
keep_active_partition_tie_breaker: When tie happens prefer partition with members of previously active (quorate) partition. This is hard-coded behavior of LMS algorithm so this setting affects only FFSplit algorithm. (off)