ocserv開源vpn部署記錄

2025-05-28 20:22 馬哥Linux運維

1、安裝ocserv

ocserv 已經在 epel 倉庫中提供了，所以可以直接透過 yum 安裝

[root@linux-node4 ~]# yum -y install epel-release
[root@linux-node4 ~]# yum -y install ocserv

2、證書製作

2.1、建立 CA

[root@linux-node4 ~]# cd /etc/ocserv
[root@linux-node4 ocserv]# mkdir CA
[root@linux-node4 ocserv]# cd CA
[root@linux-node4 CA]# cat > ca.tmpl <<EOF
cn = "pugongying"
organization = "Test Qi"
serial = 1
expiration_days = -1
ca
signing_key
cert_signing_key
crl_signing_key
EOF
[root@linux-node4 CA]# ll
total 4
-rw-r--r-- 1 root root 127 Nov 2510:42 ca.tmpl

2.2、生成CA 金鑰

[root@linux-node4 CA]# certtool --generate-privkey --outfile ca-key.pem
[root@linux-node4 CA]# ll
total 12
-rw------- 1 root root 5816 Nov 25 10:42 ca-key.pem
-rw-r--r-- 1 root root 127 Nov 25 10:42 ca.tmpl

2.3、生成CA證書

[root@linux-node4 CA]# certtool --generate-self-signed--load-privkey ca-key.pem--template ca.tmpl--outfile ca-cert.pem
[root@linux-node4 CA]# ll
total 16
-rw-r--r--1 root root 1107 Nov 2510:45 ca-cert.pem
-rw------- 1 root root 5816 Nov 2510:42 ca-key.pem
-rw-r--r-- 1 root root 127 Nov 2510:42 ca.tmpl

2.4、構建一個CA認證中心

[root@linux-node4 CA]# ip a1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128scope host
valid_lft forever preferred_lft forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5c:0e:b1 brd ff:ff:ff:ff:ff:ff
inet 192.168.56.24/24 brd 192.168.56.255scopeglobal eth0
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fe5c:eb1/64scope link
valid_lft forever preferred_lft forever
[root@linux-node4 CA]# cat > server.tmpl <<EOF
#cn 為伺服器ip或者執行該伺服器的域名
cn = "192.168.56.24"
organization = "jq"
expiration_days =36500
signing_key
encryption_key
tls_www_server
EOF
[root@linux-node4 CA]# ll
total 20
-rw-r--r-- 1 root root 1107 Nov 25 10:45 ca-cert.pem
-rw------- 1 root root 5816 Nov 25 10:42 ca-key.pem
-rw-r--r-- 1 root root 127 Nov 25 10:42 ca.tmpl
-rw-r--r-- 1 root root 163 Nov 25 10:55 server.tmpl

2.5、建立Server 金鑰

[root@linux-node4 CA]# certtool --generate-privkey --outfile server-key.pem
[root@linux-node4 CA]# ll
total 28
-rw-r--r-- 1 root root 1107 Nov 25 10:45 ca-cert.pem
-rw------- 1 root root 5816 Nov 25 10:42 ca-key.pem
-rw-r--r-- 1 root root 127 Nov 25 10:42 ca.tmpl
-rw------- 1 root root 5823 Nov 25 10:58 server-key.pem
-rw-r--r-- 1 root root 163 Nov 25 10:55 server.tmpl

2.6、建立Server 證書

[root@linux-node4 CA]# certtool --generate-certificate--load-privkey server-key.pem--load-ca-certificate ca-cert.pem--load-ca-privkey ca-key.pem--template server.tmpl--outfile server-cert.pem
[root@linux-node4 CA]# ll
total 32
-rw-r--r--1 root root 1107 Nov 2510:45 ca-cert.pem
-rw------- 1 root root 5816 Nov 2510:42 ca-key.pem
-rw-r--r-- 1 root root 127 Nov 2510:42 ca.tmpl
-rw-r--r-- 1 root root 1192 Nov 2511:00 server-cert.pem
-rw------- 1 root root 5823 Nov 2510:58 server-key.pem
-rw-r--r-- 1 root root 163 Nov 2510:55 server.tmpl

3、建立路由分組管理目錄

[root@linux-node4 ~]# mkdir -p /etc/ocserv/group
[root@linux-node4 ~]# ll /etc/ocserv/group
total 16
-rw-r--r-- 1 root root 144 Nov 25 10:11 dev
-rw-r--r-- 1 root root 145 Nov 22 15:45 front
-rw-r--r-- 1 root root 171 Nov 22 17:06 ops
-rw-r--r-- 1 root root 0 Nov 22 10:37 pre
-rw-r--r-- 1 root root 0 Nov 22 10:37 pro
-rw-r--r-- 1 root root 145 Nov 22 15:45 test
[root@linux-node4 ~]# cat /etc/ocserv/group/ops
route =192.168.5.0/255.255.255.0
route =172.17.134.0/255.255.255.0
route =172.17.2.0/255.255.255.0
route =172.17.29.0/255.255.255.0
route =172.17.172.0/255.255.255.0

  • 註釋：填寫上路由的才會走vpn，如果一個都沒有填寫預設會認為全部都走vpn，會導致上不去網。

4、修改配置檔案

備份修改配置檔案

[root@linux-node4 ~]# cp /etc/ocserv/ocserv.conf{,_bak}
[root@linux-node4 ~]# egrep -v "^$|#" /etc/ocserv/ocserv.conf
#密碼認證，選擇這個就可以通過後面的建立使用者命令直接建立使用者來登入
auth = "plain[/etc/ocserv/ocpasswd]"
#tcp埠
tcp-port = 443
#執行使用者
run-as-user = ocserv
#執行組
run-as-group = ocserv
# 分組檔案儲存目錄，檔案內容為不同的route鏈路訪問控制default-group-config = /etc/ocserv/group/Default
# 新增,用於使用者分組.default-select-group = Default
# 新增,用於使用者分組.
config-per-group = /etc/ocserv/group/
# 新增,用於使用者分組.
auto-select-group = false
socket-file = ocserv.sock
chroot-dir = /var/lib/ocserv
isolate-workers = true
# 最大使用者數量
max-clients = 1024
# 同一個使用者最多同時登陸數
max-same-clients = 10
# 3小時以內保持連線(3小時後無操作時，斷開ocserv連線)
keepalive = 10800
dpd = 90
mobile-dpd = 1800
switch-to-tcp-timeout = 25
# 證書路徑
server-cert = /etc/ocserv/CA/server-cert.pem
# 證書路徑
server-key = /etc/ocserv/CA/server-key.pem
# ca路徑
ca-cert = /etc/ocserv/CA/ca-cert.pem
tls-priorities = "NORMAL:%SERVER_PRECEDENCE:%COMPAT:-VERS-SSL3.0"
# 在進行身份驗證之前，允許客戶端保持連線的時間（秒）
auth-timeout = 240
# 身份驗證嘗試失敗後不允許客戶端重新連線的時間（秒）。
min-reauth-time = 300
max-ban-score = 0
ban-reset-time = 300
cookie-timeout = 300
deny-roaming = false
# ocserv將要求客戶機在經過這段時間後定期重新整理金鑰。設定為零以停用（請注意，如果停用重新設定金鑰，則某些客戶端將失敗）【此設定並非要求更改客戶端秘鑰】
rekey-time = 172800
rekey-method = ssl
use-occtl = true
pid-file = /var/run/ocserv.pid
device = vpns
predictable-ips = truedefault-domain = meitian-op-dev-alpha
# 分配給VPN客戶端的IP段
ipv4-network = 192.168.50.0/24
# DNS
dns = 192.168.5.1
# 預設走vpn的域名
# split-dns = elk.test.com
# split-dns = zabbix.test.com
# split-dns = jenkins.test.com
# split-dns = *.test.com
ping-leases = false
cisco-client-compat = true
dtls-legacy = true
user-profile = profile.xml

原配置檔案備份

（/etc/ocserv/ocserv.conf_bak — 此步驟可忽略不看）

[root@linux-node4 ~]# cat /etc/ocserv/ocserv.conf_bak
### The following directives do not change with server reload.## User authentication method. To require multiple methods to be
# used for the user to login, add multiple auth directives. The values
# in the 'auth' directive are AND composed (if multiple all must
# succeed).
# Available options: certificate, plain, pam, radius, gssapi.
# Note that authentication methods utilizing passwords cannot be
# combined (e.g., the plain, pam or radius methods).
#
# certificate:
# This indicates that all connecting users must present a certificate.
# The username and user group will be then extracted from it (see
# cert-user-oid and cert-group-oid). The certificate to be accepted
# it must be signed by the CA certificate as specified in 'ca-cert' and
# it must not be listed in the CRL, as specified by the 'crl' option.
#
# pam[gid-min=1000]:
# This enabled PAM authentication of the user. The gid-min option is used
# by auto-select-group option, in order to select the minimum valid group ID.
#
# plain[passwd=/etc/ocserv/ocpasswd,otp=/etc/ocserv/users.otp]
# The plain option requires specifying a password file which contains
# entries of the following format.
# "username:groupname1,groupname2:encoded-password"
# One entry must be listed per line, and 'ocpasswd' should be used
# to generate password entries. The 'otp' suboption allows to specify
# an oath password file to be used for one time passwords; the format of
# the file is described in https://code.google.com/p/mod-authn-otp/wiki/UsersFile#
# radius[config=/etc/radiusclient/radiusclient.conf,groupconfig=true,nas-identifier=name,override-interim-updates=false]:
# The radius option requires specifying freeradius-client configuration
# file. If the groupconfig option is set, then config-per-user will be overriden,
# and all configuration will be read from radius. The 'override-interim-updates' if set to
# true will ignore Acct-Interim-Interval from the server and 'stats-report-time' will be considered.
#
# gssapi[keytab=/etc/key.tab,require-local-user-map=true,tgt-freshness-time=900]
# The gssapi option allows to use authentication methods supported by GSSAPI,
# such as Kerberos tickets with ocserv. It should be best used as an alternative
# to PAM (i.e., have pam in auth and gssapi in enable-auth), to allow users with
# tickets and without tickets to login. The default value for require-local-user-map
# is true. The 'tgt-freshness-time' if set, it would require the TGT tickets presented
# to have been issued within the provided number of seconds. That option is used to
# restrict logins even if the KDC provides long time TGT tickets.
auth = "pam"
#auth = "pam[gid-min=1000]"
#auth = "plain[passwd=./sample.passwd,otp=./sample.otp]"
#auth = "certificate"
#auth = "radius[config=/etc/radiusclient/radiusclient.conf,groupconfig=true]"
# Specify alternative authentication methods that are sufficient
# for authentication. That is, if set, any of the methods enabled
# will be sufficient to login, irrespective of the main 'auth' entries.
# When multiple options are present, they are OR composed (any of them
# succeeding allows login).
#enable-auth = "certificate"
#enable-auth = "gssapi"
#enable-auth = "gssapi[keytab=/etc/key.tab,require-local-user-map=true,tgt-freshness-time=900]"
# Accounting methods available:
# radius: can be combined with any authentication method, it provides
# radius accounting to available users (see also stats-report-time).
#
# pam: can be combined with any authentication method, it provides
# a validation of the connecting user's name using PAM. It is
# superfluous to use this method when authentication is already
# PAM.
#
# Only one accounting method can be specified.
#acct = "radius[config=/etc/radiusclient/radiusclient.conf]"
# Use listen-host to limit to specific IPs or to the IPs of a provided
# hostname.
#listen-host = [IP|HOSTNAME]
# When the server has a dynamic DNS address (that may change),
# should set that to true to ask the client to resolve again on
# reconnects.
#listen-host-is-dyndns = true
# TCP and UDP port numbertcp-port = 443
udp-port = 443
# Accept connections using a socket file. It accepts HTTP
# connections (i.e., without SSL/TLS unlike its TCP counterpart),
# and uses it as the primary channel. That option cannot be
# combined with certificate authentication.
#listen-clear-file = /var/run/ocserv-conn.socket
# The user the worker processes will be run as. It should be
＃將作為工作程序執行的使用者。
# unique (no other services run as this user).
# 唯一（沒有其他服務以此使用者身份執行）。run-as-user = ocservrun-as-group = ocserv
# socket file used for IPC with occtl. You only need to set that,
# if you use more than a single servers.
#occtl-socket-file = /var/run/occtl.socket
# socket file used for server IPC (worker-main), will be appended with .PID
# It must be accessible within the chroot environment (if any), so it is best
# specified relatively to the chroot directory.socket-file = ocserv.sock
# The default server directory. Does not require any devices present.chroot-dir = /var/lib/ocserv
### All configuration options below this line are reloaded on a SIGHUP.
### The options above, will remain unchanged. Note however, that the
### server-cert, server-key, dh-params and ca-cert options will be reloaded
### if the provided file changes, on server reload. That allows certificate
### rotation, but requires the server key to remain the same for seamless
### operation. If the server key changes on reload, there may be connection
### failures during the reloading time.
# Whether to enable seccomp/Linux namespaces worker isolation. That restricts the number of
# system calls allowed to a worker process, in order to reduce damage from a
# bug in the worker process. It is available on Linux systems at a performance cost.
# The performance cost is roughly 2% overhead at transfer time (tested on a Linux 3.17.8).
# Note however, that process isolation is restricted to the specific libc versions
# the isolation was tested at. If you get random failures on worker processes, try
# disabling that option and report the failures you, along with system and debugging
# information at: https://gitlab.com/ocserv/ocserv/issuesisolate-workers = true
# A banner to be displayed on clients
#banner = "Welcome"
# Limit the number of clients. Unset or set to zero for unlimited.
#max-clients = 1024
max-clients = 16
# Limit the number of identical clients (i.e., users connecting
# multiple times). Unset or set to zero for unlimited.max-same-clients = 2
# Limit the number of client connections to one every X milliseconds
# (X is the provided value). Set to zero for no limit.
#rate-limit-ms = 100
# Stats report time. The number of seconds after which each
# worker process will report its usage statistics (number of
# bytes transferred etc). This is useful when accounting like
# radius is in use.
#stats-report-time = 360
# Keepalive in secondskeepalive = 32400
# Dead peer detection in seconds.
# Note that when the client is behind a NAT this value
# needs to be short enough to prevent the NAT disassociating
# his UDP session from the port number. Otherwise the client
# could have his UDP connection stalled, for several minutes.dpd = 90
# Dead peer detection for mobile clients. That needs to
# be higher to prevent such clients being awaken too
# often by the DPD messages, and save battery.
# The mobile clients are distinguished from the header
# 'X-AnyConnect-Identifier-DeviceType'.mobile-dpd = 1800
# If using DTLS, and no UDP traffic is received for this
# many seconds, attempt to send future traffic over the TCP
# connection instead, in an attempt to wake up the client
# in the case that there is a NAT and the UDP translation
# was deleted. If this is unset, do not attempt to use this
# recovery mechanism.switch-to-tcp-timeout = 25
# MTU discovery (DPD must be enabled)try-mtu-discovery = false
# The key and the certificates of the server
# The key may be a file, or any URL supported by GnuTLS (e.g.,
# tpmkey:uuid=xxxxxxx-xxxx-xxxx-xxxx-xxxxxxxx;storage=user
# or pkcs11:object=my-vpn-key;object-type=private)
#
# The server-cert file may contain a single certificate, or
# a sorted certificate chain.
#
# There may be multiple server-cert and server-key directives,
# but each key should correspond to the preceding certificate.
# The certificate files will be reloaded when changed allowing for in-place
# certificate renewal (they are checked and reloaded periodically;
# a SIGHUP signal to main server will force reload).
server-cert = /etc/pki/ocserv/public/server.crtserver-key = /etc/pki/ocserv/private/server.key
# Diffie-Hellman parameters. Only needed if you require support
# for the DHE ciphersuites (by default this server supports ECDHE).
# Can be generated using:
# certtool --generate-dh-params --outfile /path/to/dh.pem
#dh-params = /path/to/dh.pem
# If you have a certificate from a CA that provides an OCSP
# service you may provide a fresh OCSP status response within
# the TLS handshake. That will prevent the client from connecting
# independently on the OCSP server.
# You can update this response periodically using:
# ocsptool --ask --load-cert=your_cert --load-issuer=your_ca --outfile response
# Make sure that you replace the following file in an atomic way.
#ocsp-response = /path/to/ocsp.der
# In case PKCS #11, TPM or encrypted keys are used the PINs should be available
# in files. The srk-pin-file is applicable to TPM keys only, and is the
# storage root key.
#pin-file = /path/to/pin.txt
#srk-pin-file = /path/to/srkpin.txt
# The password or PIN needed to unlock the key in server-key file.
# Only needed if the file is encrypted or a PKCS #11 object. This
# is an alternative method to pin-file.
#key-pin = 1234
# The SRK PIN for TPM.
# This is an alternative method to srk-pin-file.
#srk-pin = 1234
# The Certificate Authority that will be used to verify
# client certificates (public keys) if certificate authentication
# is set.ca-cert = /etc/pki/ocserv/cacerts/ca.crt
# The object identifier that will be used to read the user ID in the client
# certificate. The object identifier should be part of the certificate's DN# Useful OIDs are:
# CN = 2.5.4.3, UID = 0.9.2342.19200300.100.1.1
cert-user-oid = 0.9.2342.19200300.100.1.1
# The object identifier that will be used to read the user group in the
# client certificate. The object identifier should be part of the certificate's# DN. Useful OIDs are:
# OU (organizational unit) = 2.5.4.11
#cert-group-oid = 2.5.4.11
# The revocation list of the certificates issued by the 'ca-cert' above.
# See the manual to generate an empty CRL initially. The CRL will be reloaded
# periodically when ocserv detects a change in the file. To force a reload use
# SIGHUP.
#crl = /path/to/crl.pem
# Uncomment this to enable compression negotiation (LZS, LZ4).
#compression = true
# Set the minimum size under which a packet will not be compressed.
# That is to allow low-latency for VoIP packets. The default size
# is 256 bytes. Modify it if the clients typically use compression
# as well of VoIP with codecs that exceed the default value.
#no-compress-limit = 256
# GnuTLS priority string; note that SSL 3.0 is disabled by default
# as there are no openconnect (and possibly anyconnect clients) using
# that protocol. The string below does not enforce perfect forward
# secrecy, in order to be compatible with legacy clients.
#
# Note that the most performant ciphersuites are the moment are the ones
# involving AES-GCM. These are very fast in x86 and x86-64 hardware, and
# in addition require no padding, thus taking full advantage of the MTU.
# For that to be taken advantage of, the openconnect client must be
# used, and the server must be compiled against GnuTLS 3.2.7 or later.
# Use "gnutls-cli --benchmark-tls-ciphers", to see the performance
# difference with AES_128_CBC_SHA1 (the default for anyconnect clients)
# in your system.
#tls-priorities = "NORMAL:%SERVER_PRECEDENCE:%COMPAT:-VERS-SSL3.0"
tls-priorities = "NORMAL:%SERVER_PRECEDENCE:%COMPAT:-VERS-SSL3.0"
# More combinations in priority strings are available, check
# http://gnutls.org/manual/html_node/Priority-Strings.html
# E.g., the string below enforces perfect forward secrecy (PFS)
# on the main channel.
#tls-priorities = "NORMAL:%SERVER_PRECEDENCE:%COMPAT:-RSA:-VERS-SSL3.0:-ARCFOUR-128"
# That option requires the established DTLS channel to use the same
# cipher as the primary TLS channel. This cannot be combined with
# listen-clear-file since the ciphersuite information is not available
# in that configuration. Note also, that this option implies that
# dtls-legacy option is false; this option cannot be enforced
# in the legacy/compat protocol.
#match-tls-dtls-ciphers = true
# The time (in seconds) that a client is allowed to stay connected prior
# to authenticationauth-timeout = 240
# The time (in seconds) that a client is allowed to stay idle (no traffic)
# before being disconnected. Unset to disable.
#idle-timeout = 1200
# The time (in seconds) that a client is allowed to stay connected
# Unset to disable.
#session-timeout = 86400
# The time (in seconds) that a mobile client is allowed to stay idle (no
# traffic) before being disconnected. Unset to disable.
#mobile-idle-timeout = 2400
# The time (in seconds) that a client is not allowed to reconnect after
# a failed authentication attempt.min-reauth-time = 300
# Banning clients in ocserv works with a point system. IP addresses
# that get a score over that configured number are banned for
# min-reauth-time seconds. By default a wrong password attempt is 10 points,
# a KKDCP POST is 1 point, and a connection is 1 point. Note that
# due to difference processes being involved the count of points
# will not be real-time precise.
#
# Score banning cannot be reliably used when receiving proxied connections
# locally from an HTTP server (i.e., when listen-clear-file is used).
#
# Set to zero to disable.max-ban-score = 50
# The time (in seconds) that all score kept for a client is reset.ban-reset-time = 300
# In case you'd like to change the default points.
#ban-points-wrong-password = 10
#ban-points-connection = 1
#ban-points-kkdcp = 1
# Cookie timeout (in seconds)
# Once a client is authenticated he's provided a cookie with
# which he can reconnect. That cookie will be invalided if not
# used within this timeout value. On a user disconnection, that
# cookie will also be active for this time amount prior to be
# invalid. That should allow a reasonable amount of time for roaming
# between different networks.cookie-timeout = 300
# If this is enabled (not recommended) the cookies will stay
# valid even after a user manually disconnects, and until they
# expire. This may improve roaming with some broken clients.
#persistent-cookies = true
# Whether roaming is allowed, i.e., if true a cookie is
# restricted to a single IP address and cannot be re-used
# from a different IP.deny-roaming = false
# ReKey time (in seconds)
# ocserv will ask the client to refresh keys periodically once
# this amount of seconds is elapsed. Set to zero to disable (note
# that, some clients fail if rekey is disabled).rekey-time = 172800
# ReKey method
# Valid options: ssl, new-tunnel
# ssl: Will perform an efficient rehandshake on the channel allowing
# a seamless connection during rekey.
# new-tunnel: Will instruct the client to discard and re-establish the channel.
# Use this option only if the connecting clients have issues with the ssl
# option.rekey-method = ssl
# Script to call when a client connects and obtains an IP.
# The following parameters are passed on the environment.
# REASON, USERNAME, GROUPNAME, HOSTNAME (the hostname selected by client),
# DEVICE, IP_REAL (the real IP of the client), IP_REAL_LOCAL (the local
# interface IP the client connected), IP_LOCAL (the local IP
# in the P-t-P connection), IP_REMOTE (the VPN IP of the client),
# IPV6_LOCAL (the IPv6 local address if there are both IPv4 and IPv6
# assigned), IPV6_REMOTE (the IPv6 remote address), IPV6_PREFIX, and
# ID (a unique numeric ID); REASON may be "connect" or "disconnect".
# In addition the following variables OCSERV_ROUTES (the applied routes for this
# client), OCSERV_NO_ROUTES, OCSERV_DNS (the DNS servers for this client),
# will contain a space separated list of routes or DNS servers. A version
# of these variables with the 4 or 6 suffix will contain only the IPv4 or
# IPv6 values.
# The disconnect script will receive the additional values: STATS_BYTES_IN,
# STATS_BYTES_OUT, STATS_DURATION that contain a 64-bit counter of the bytes
# output from the tun device, and the duration of the session in seconds.
#connect-script = /usr/bin/ocserv-script
#disconnect-script = /usr/bin/ocserv-script
# UTMP
# Register the connected clients to utmp. This will allow viewing
# the connected clients using the command 'who'.
#use-utmp = true
# Whether to enable support for the occtl tool (i.e., either through D-BUS,
# or via a unix socket).use-occtl = true
# PID file. It can be overriden in the command line.pid-file = /var/run/ocserv.pid
# Set the protocol-defined priority (SO_PRIORITY) for packets to
# be sent. That is a number from 0 to 6 with 0 being the lowest
# priority. Alternatively this can be used to set the IP Type-
# Of-Service, by setting it to a hexadecimal number (e.g., 0x20).
# This can be set per user/group or globally.
#net-priority = 3
# Set the VPN worker process into a specific cgroup. This is Linux
# specific and can be set per user/group or globally.
#cgroup = "cpuset,cpu:test"
#
# Network settings
#
# The name to use for the tun devicedevice = vpns
# Whether the generated IPs will be predictable, i.e., IP stays the
# same for the same user when possible.predictable-ips = true
# The default domain to be advertiseddefault-domain = example.com
# The pool of addresses that leases will be given from. If the leases
# are given via Radius, or via the explicit-ip? per-user config option then
# these network values should contain a network with at least a single
# address that will remain under the full control of ocserv (that is
# to be able to assign the local part of the tun device address).
#ipv4-network = 192.168.1.0
#ipv4-netmask = 255.255.255.0
# An alternative way of specifying the network:
#ipv4-network = 192.168.1.0/24
# The IPv6 subnet that leases will be given from.
#ipv6-network = fda9:4efe:7e3b:03ea::/64
# Specify the size of the network to provide to clients. It is
# generally recommended to provide clients with a /64 network in
# IPv6, but any subnet may be specified. To provide clients only
# with a single IP use the prefix 128.
#ipv6-subnet-prefix = 128
#ipv6-subnet-prefix = 64
# Whether to tunnel all DNS queries via the VPN. This is the default
＃是否透過VPN隧道傳輸所有DNS查詢。 這是預設值
# when a default route is set.
＃設定預設路由時。
#tunnel-all-dns = true
# The advertized DNS server. Use multiple lines for
# multiple servers.
# dns = fc00::4be0
#dns = 192.168.1.2
# The NBNS server (if any)
#nbns = 192.168.1.3
# The domains over which the provided DNS should be used. Use
# multiple lines for multiple domains.
# 使用提供的DNS域，可以寫多個（寫上的域名會走dns解析到dns伺服器，不寫則不會走dns）
#split-dns = example.com
# Prior to leasing any IP from the pool ping it to verify that
# it is not in use by another (unrelated to this server) host.
# Only set to true, if there can be occupied addresses in the
# IP range for leases.ping-leases = false
# Use this option to enforce an MTU value to the incoming
# connections. Unset to use the default MTU of the TUN device.
#mtu = 1420
# Unset to enable bandwidth restrictions (in bytes/sec). The
# setting here is global, but can also be set per user or per group.
#rx-data-per-sec = 40000
#tx-data-per-sec = 40000
# The number of packets (of MTU size) that are available in
# the output buffer. The default is low to improve latency.
# Setting it higher will improve throughput.
#output-buffer = 10
# Routes to be forwarded to the client. If you need the
# client to forward routes to the server, you may use the
# config-per-user/group or even connect and disconnect scripts.
#
# To set the server as the default gateway for the client just
# comment out all routes from the server, or use the special keyword
# 'default'.
#route = 10.10.10.0/255.255.255.0
#route = 192.168.0.0/255.255.0.0
#route = fef4:db8:1000:1001::/64
# Subsets of the routes above that will not be routed by
# the server.
#no-route = 192.168.5.0/255.255.255.0
# If set, the script /usr/bin/ocserv-fw will be called to restrict
# the user to its allowed routes and prevent him from accessing
# any other routes. In case of defaultroute, the no-routes are restricted.
# All the routes applied by ocserv can be reverted using /usr/bin/ocserv-fw
# --removeall. This option can be set globally or in the per-user configuration.
#restrict-user-to-routes = true
# When set to true, all client's iroutes are made visible to all
# connecting clients except for the ones offering them. This option
# only makes sense if config-per-user is set.
#expose-iroutes = true
# Groups that a client is allowed to select from.
# A client may belong in multiple groups, and in certain use-cases
# it is needed to switch between them. For these cases the client can
# select prior to authentication. Add multiple entries for multiple groups.
# The group may be followed by a user-friendly name in brackets.
#select-group = group1
#select-group = group2[My special group]
# The name of the (virtual) group that if selected it would assign the user
# to its default group.
#default-select-group = DEFAULT
# Instead of specifying manually all the allowed groups, you may instruct
# ocserv to scan all available groups and include the full list.
#auto-select-group = true
# Configuration files that will be applied per user connection or
# per group. Each file name on these directories must match the username
# or the groupname.
# The options allowed in the configuration files are dns, nbns,
# ipv?-network, ipv4-netmask, rx/tx-per-sec, iroute, route, no-route,
# explicit-ipv4, explicit-ipv6, net-priority, deny-roaming, no-udp,
# user-profile, cgroup, stats-report-time, and session-timeout.
#
# Note that the 'iroute' option allows to add routes on the server
# based on a user or group. The syntax depends on the input accepted
# by the commands route-add-cmd and route-del-cmd (see below). The no-udp
# is a boolean option (e.g., no-udp = true), and will prevent a UDP session
# for that specific user or group.
#config-per-user = /etc/ocserv/config-per-user/
#config-per-group = /etc/ocserv/config-per-group/
# When config-per-xxx is specified and there is no group or user that
# matches, then utilize the following configuration.
#default-user-config = /etc/ocserv/defaults/user.conf
#default-group-config = /etc/ocserv/defaults/group.conf
# The system command to use to setup a route. %{R} will be replaced with the
# route/mask and %{D} with the (tun) device.
#
# The following example is from linux systems. %R should be something
# like 192.168.2.0/24 (the argument of iroute).
#route-add-cmd = "ip route add %{R} dev %{D}"
#route-del-cmd = "ip route delete %{R} dev %{D}"
# This option allows to forward a proxy. The special keywords '%{U}'
# and '%{G}', if present will be replaced by the username and group name.
#proxy-url = http://example.com/
#proxy-url = http://example.com/%{U}/
# This option allows you to specify a URL location where a client can
# post using MS-KKDCP, and the message will be forwarded to the provided
# KDC server. That is a translation URL between HTTP and Kerberos.
# In MIT kerberos you'll need to add in realms:
# EXAMPLE.COM = {
# kdc = https://ocserv.example.com/kerberos
# http_anchors = FILE:/etc/ocserv-ca.pem
# }
# This option is available if ocserv is compiled with GSSAPI support.
#kkdcp = SERVER-PATH KERBEROS-REALM PROTOCOL@SERVER:PORT
#kkdcp = /kerberos EXAMPLE.COM [email protected]:88
#kkdcp = /kerberos-tcp EXAMPLE.COM [email protected]:88
#
# The following options are for (experimental) AnyConnect client
# compatibility.
# This option will enable the pre-draft-DTLS version of DTLS, and
# will not require clients to present their certificate on every TLS
# connection. It must be set to true to support legacy CISCO clients
# and openconnect clients < 7.08. When set to true, it implies dtls-legacy = true.cisco-client-compat = true
# This option allows to disable the DTLS-PSK negotiation (enabled by default).
# The DTLS-PSK negotiation was introduced in ocserv 0.11.5 to deprecate
# the pre-draft-DTLS negotiation inherited from AnyConnect. It allows the
# DTLS channel to negotiate its ciphers and the DTLS protocol version.
#dtls-psk = false
# This option allows to disable the legacy DTLS negotiation (enabled by default,
# but that may change in the future).
# The legacy DTLS uses a pre-draft version of the DTLS protocol and was
# from AnyConnect protocol. It has several limitations, that are addressed
# by the dtls-psk protocol supported by openconnect 7.08+.dtls-legacy = true
# Client profile xml. A sample file exists in doc/profile.xml.
# It is required by some of the CISCO clients.
# This file must be accessible from inside the worker's chroot. user-profile = profile.xml
#Advanced options
# Option to allow sending arbitrary custom headers to the client after
# authentication and prior to VPN tunnel establishment. You shouldn't
# need to use this option normally; if you do and you think that
# this may help others, please send your settings and reason to
# the openconnect mailing list. The special keywords '%{U}'
# and '%{G}', if present will be replaced by the username and group name.
#custom-header = "X-My-Header: hi there"

5、重啟ocserv

【撥通vpn後會出現網絡卡vpns0資訊】

[root@linux-node4 ~]# ip a1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5c:0e:b1 brd ff:ff:ff:ff:ff:ff
inet 192.168.56.24/24 brd 192.168.56.255 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fe5c:eb1/64 scope link
valid_lft forever preferred_lft forever
[root@linux-node4 ~]# systemctl restart ocserv

#撥通vpn後會出現網絡卡vpns0資訊

[root@linux-node4~]# ifconfig
eth0:flags=4163<UP,BROADCAST,RUNNING,MULTICAST>mtu1500
inet192.168.56.24netmask255.255.255.0broadcast192.168.56.255
inet6fe80::20c:29ff:fe5c:eb1prefixlen64scopeid0x20<link>
ether00:0c:29:5c:0e:b1txqueuelen1000(Ethernet)
RXpackets45601bytes31267946(29.8MiB)
RXerrors0dropped0overruns0frame0
TXpackets11299bytes1249859(1.1MiB)
TXerrors0dropped0overruns0carrier0collisions0
lo:flags=73<UP,LOOPBACK,RUNNING>mtu65536
inet127.0.0.1netmask255.0.0.0
inet6::1prefixlen128scopeid0x10<host>
looptxqueuelen1(LocalLoopback)
RXpackets0bytes0(0.0B)
RXerrors0dropped0overruns0frame0
TXpackets0bytes0(0.0B)
TXerrors0dropped0overruns0carrier0collisions0
vpns0:flags=81<UP,POINTOPOINT,RUNNING>mtu1472
inet192.168.5.1netmask255.255.255.255destination192.168.5.90
inet6fe80::698d:d063:ff73:4260prefixlen64scopeid0x20<link>
unspec00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00txqueuelen500(UNSPEC)
RXpackets125bytes14790(14.4KiB)
RXerrors0dropped0overruns0frame0
TXpackets17bytes1349(1.3KiB)
TXerrors0dropped0overruns0carrier0collisions0
[root@linux-node4~]# netstat -lntup|grep 443
tcp000.0.0.0:4430.0.0.0:*LISTEN21069/ocserv-main
tcp600:::443:::*LISTEN21069/ocserv-main

6、使用者管理

建立一個登陸用的使用者名稱與密碼

[root@linux-node4 ~]# ocpasswd -c /etc/ocserv/ocpasswd test
Enter password: 123456
Re-enter password: 123456
[root@linux-node4 ~]# cat /etc/ocserv/ocpasswd
test:*:$5$TtT/7WUcqH6vV/IA$SSs5yKVjVPS4qO96dvNRtkvcnX9dAPfLNfpjCUYOTg8

7、配置防火牆

network 192.169.5.0/24 (這個來自於)/etc/ocserv/ocserv.conf 中的 ipv4-network = 192.168.5.0/24

ocserv WAN interface is eth0

7.1、修改核心配置

[root@linux-node4 ~]# cp /etc/sysctl.conf{,_bak}
[root@linux-node4 ~]# vim /etc/sysctl.conf
# sysctl settings are defined through files in
# /usr/lib/sysctl.d/, /run/sysctl.d/, and /etc/sysctl.d/.
#
# Vendors settings live in /usr/lib/sysctl.d/.
# To override a whole file, create a new file with the same in
# /etc/sysctl.d/ and put new settings there. To override
# only specific settings, add a file with a lexically later
# name in /etc/sysctl.d/ and put new settings there.
#
# For more information, see sysctl.conf(5) and sysctl.d(5).
net.ipv6.conf.all.disable_ipv6 = 1
net.ipv6.conf.default.disable_ipv6 = 1
net.ipv6.conf.lo.disable_ipv6 = 1
vm.swappiness = 0
net.ipv4.neigh.default.gc_stale_time=120
# see details in https://help.aliyun.com/knowledge_detail/39428.html
# Accept IPv6 advertisements when forwarding is enabled
# # Protect from IP Spoofing
net.ipv4.conf.all.rp_filter=1
net.ipv4.conf.default.rp_filter=1
# Ignore ICMP broadcast requests
net.ipv4.icmp_echo_ignore_broadcasts = 1
# Protect from bad icmp error messages
net.ipv4.icmp_ignore_bogus_error_responses = 1
# Disable source packet routing
net.ipv4.conf.all.accept_source_route = 0
net.ipv6.conf.all.accept_source_route = 0
net.ipv4.conf.default.accept_source_route = 0
net.ipv6.conf.default.accept_source_route = 0
# Turn on exec shield
kernel.exec-shield = 1
kernel.randomize_va_space = 1
net.ipv4.conf.default.arp_announce = 2
net.ipv4.conf.lo.arp_announce=2
net.ipv4.conf.all.arp_announce=2
# see details in https://help.aliyun.com/knowledge_detail/41334.html
net.ipv4.tcp_max_tw_buckets = 5000
# Block SYN attacks
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_max_syn_backlog = 1024
net.ipv4.tcp_synack_retries = 2
net.ipv4.tcp_syn_retries = 5
# Ignore send redirects
net.ipv4.conf.all.send_redirects = 0
net.ipv4.conf.default.send_redirects = 0
# Ignore ICMP redirects
net.ipv4.conf.all.accept_redirects = 0
net.ipv6.conf.all.accept_redirects = 0
net.ipv4.conf.default.accept_redirects = 0
net.ipv6.conf.default.accept_redirects = 0
net.ipv4.conf.all.secure_redirects = 0
net.ipv4.conf.default.secure_redirects = 0
net.ipv6.conf.all.accept_ra = 2
net.ipv6.conf.eth0.accept_ra = 2
net.ipv4.ip_forward = 1
kernel.sysrq=1
# 允許將TIME-WAIT sockets重新用於新的TCP連線，預設為0，表示關閉(此引數解決：Cannot assign requested address【客戶端連線失敗問題】)
net.ipv4.tcp_tw_reuse = 1
# 表示開啟TCP連線中TIME-WAIT sockets的快速回收，預設為0，表示關閉(此引數解決：Cannot assign requested address【客戶端連線失敗問題】)
net.ipv4.tcp_tw_recycle = 1

7.2、生效配置

[root@linux-node4 ~]# sysctl -p

7.3、關閉centos原有防火牆

[root@linux-node4 ~]# systemctl stop firewalld
[root@linux-node4 ~]# systemctl mask firewalld

7.4、安裝新的iptables-services

[root@linux-node4 ~]# yum install iptables-services
[root@linux-node4 ~]# service iptables start

7.5、配置iptables

[root@linux-node4~]# vim /etc/sysconfig/iptables
# sample configuration for iptables service
# you can edit this manually or use system-config-firewall
# please do not ask us to add additional ports/services to this default configuration
#*filter
#:INPUT ACCEPT [0:0]
#:FORWARD ACCEPT [0:0]
#:OUTPUT ACCEPT [0:0]
#-A INPUT -m state --state RELATED,ESTABLISHED -j ACCEPT
#-A INPUT -p icmp -j ACCEPT
#-A INPUT -i lo -j ACCEPT
#-A INPUT -p tcp -m state --state NEW -m tcp --dport 22 -j ACCEPT
#-A INPUT -j REJECT --reject-with icmp-host-prohibited
#-A FORWARD -j REJECT --reject-with icmp-host-prohibited
#COMMIT
#*nat
:INPUTACCEPT [0:0]
:PREROUTINGACCEPT [0:0]
:OUTPUTACCEPT [0:0]
:POSTROUTINGACCEPT [0:0]
# Generic NAT for LAN Network 192.168.5.0/24
-APOSTROUTING-s192.168.5.0/24-oeth0-jMASQUERADE
COMMIT
*mangle
:PREROUTINGACCEPT [0:0]
:INPUTACCEPT [0:0]
:FORWARDACCEPT [0:0]
:OUTPUTACCEPT [0:0]
:POSTROUTINGACCEPT [0:0]
COMMIT
*filter
:INPUTACCEPT [0:0]
:FORWARDACCEPT [0:0]
:OUTPUTACCEPT [0:0]
# START INPUT RULES
# Stateful Rule - INPUT-A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
# ACCEPT traffic from Loopback interface
-AINPUT-ilo-jACCEPT
# ACCEPT SSH from LAN-A INPUT -p tcp -m tcp -i eth0 --dport 22 -j ACCEPT
# ACCEPT DHCP from LAN-A INPUT -p udp -m udp -i eth1 --dport 67:68 -j ACCEPT
# ACCEPT Webmin from LAN (Optional, only for Webmin users)-A INPUT -p tcp -m tcp -i eth0 --dport 10000 -j ACCEPT
# ACCEPT DNS UDP From LAN-A INPUT -p udp -m udp -i eth0 --dport 53 -j ACCEPT
# ACCEPT DNS TCP From LAN-A INPUT -p tcp -m tcp -i eth0 --dport 53 -j ACCEPT
# ACCEPT ping from LAN-A INPUT -p icmp --icmp-type echo-request -i eth1 -j ACCEPT
# ACCEPT OpenConnect TCP From WAN-A INPUT -p tcp -m tcp -i eth0 --dport 443 -j ACCEPT
# ACCEPT OpenConnect UPD From WAN-A INPUT -p udp -m udp -i eth0 --dport 443 -j ACCEPT
# DROP wan traffic-A INPUT -i eth0 -j DROP
# LOG LAN-A INPUT -i eth1 -j LOG --log-prefix "IPTABLES-LOG-INPUT-LAN:" --log-level 4
# ACCEPT LAN traffic - Learning rule - Should be changed to DROP once custom rules are created.-A INPUT -i eth1 -j ACCEPT
# LAST RULE - DROP all traffic-A INPUT -j DROP
# END INPUT RULES
# START FORWARD RULES
# Stateful Rule - FORWARD-A FORWARD -m state --state ESTABLISHED,RELATED -j ACCEPT
# ACCEPT LAN to WAN-A FORWARD -s 192.168.5.0/24 -j ACCEPT
# LOG Forwarded traffic-A FORWARD -j LOG --log-prefix "IPTABLES-LOG-FORWARD:" --log-level 4
# LAST RULE - ACCEPT all traffic - Should be changed to DROP once custom rules are created.-A FORWARD -j ACCEPT
# END FORWARD RULES
# START OUTPUT RULES
# Stateful Rule - OUTPUT-A OUTPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
# LOG Outgoing traffic-A OUTPUT -j LOG --log-prefix "IPTABLES-LOG-OUTPUT:" --log-level 4
# LAST RULE - ACCEPT all traffic - Should be changed to DROP once custom rules are created.-A OUTPUT -j ACCEPT
# END OUTPUT RULES
COMMIT

7.6、firewalld防火牆規則

# Generated by iptables-save v1.4.21 on Fri Apr 17 09:43:11 2020
*filter
:INPUT ACCEPT [1061568:1198147492]
:FORWARD ACCEPT [1325788:901414323]
:OUTPUT ACCEPT [696011:147931358]
COMMIT
# Completed on Fri Apr 17 09:43:11 2020
# Generated by iptables-save v1.4.21 on Fri Apr 17 09:43:11 2020
*nat
:PREROUTING ACCEPT [28781:1919648]
:INPUT ACCEPT [26802:1608227]
:OUTPUT ACCEPT [3773:270622]
:POSTROUTING ACCEPT [5:468]-A POSTROUTING -o eth0 -j MASQUERADE
COMMIT
# Completed on Fri Apr 17 09:43:11 2020

7.7、生效配置【重啟機器後需要重新生效一下】

[root@linux-node4 ~]# iptables-restore < /etc/sysconfig/iptables

8.ocserv管理

一、新增使用者

ocpasswd -c /etc/ocserv/ocpasswd 【使用者名稱】
二、新增使用者至某個分組

ocpasswd -c /etc/ocserv/ocpasswd -g 【分組名稱】 【使用者名稱】
三、鎖定使用者

ocpasswd -c /etc/ocserv/ocpasswd -l 【使用者名稱】
四、解鎖使用者

ocpasswd -c /etc/ocserv/ocpasswd -u 【使用者名稱】
五、刪除使用者

ocpasswd -c /etc/ocserv/ocpasswd -d 【使用者名稱】
六、檢視當前服務執行狀態:

occtl -n show status
七、檢視當前線上使用者詳情:

occtl -n show users
八、踢掉當前線上使用者:透過使用者名稱:

occtl disconnect user 【使用者名稱】
九、踢掉當前線上使用者：透過id:

occtl disconnect id 【id號】
連結：https://www.cnblogs.com/pgyLang/p/16349241.html
（版權歸原作者所有，侵刪）


相關文章