ssh-broker-config — Tectia Connection Broker configuration file format
The Connection Broker configuration file ssh-broker-config.xml
is used by Tectia Client and
ConnectSecure on Unix and Windows, and additionally by the Tectia client tools on IBM z/OS and Linux for
IBM System z. The Connection Broker configuration file must be a valid XML file that follows the
ssh-broker-ng-config-1.dtd
document type definition.
The SOCKS Proxy configuration file ssh-socks-proxy-config.xml
is used
by the Tectia SOCKS Proxy on z/OS. It must be a valid XML file that follows the
ssh-broker-ng-config-1.dtd
document type definition.
The Connection Broker reads three configuration files (if all are available):
The ssh-broker-config-default.xml
file is read first. It holds
the factory default settings. It is not recommended to edit the file, but you can use it
to view the default settings.
This file must be available and correctly formatted for the Connection Broker to start.
Next, the Connection Broker reads the global configuration file. The settings in the global configuration file override the default settings.
If the global configuration file is missing or malformed, the Connection Broker will start normally, and will read the user-specific configuration file, instead. A malformed global configuration file is ignored and the default settings or user-specific settings, if they exist, are used instead.
Last, the Connection Broker reads the user-specific configuration file, if it is available. The settings in the user-specific configuration file override the settings in the global configuration file, with the following exceptions:
The following settings from the user-specific configuration are combined with the settings of the global configuration file:
In general
element, the key-stores
,
cert-validation
and file-access-control
settings
In profiles
element, all settings
In static-tunnels
element, all settings.
If a connection profile with the same name has been defined in both the global configuration file and user-specific configuration file, the latter one is used.
If the filter-engine
settings have been defined in the global
configuration file, and the file is valid (not malformed), those settings are used,
and any filter-engine
settings made in the user-specific
configuration file are ignored.
If the user-specific configuration file is missing, the Connection Broker will start using the previously read configuration files. However, if a user-specific configuration exists but is malformed, the Connection Broker will not start at all.
On z/OS, the default configuration file locations are as follows:
the default configuration:
/opt/tectia/etc/ssh-tectia/auxdata/ssh-broker-ng/ssh-broker-config-default.xml
the global configuration: /opt/tectia/etc/ssh-broker-config.xml
the user-specific configuration:
$HOME/.ssh2/ssh-broker-config.xml
the XML DTD:
/opt/tectia/etc/ssh-tectia/auxdata/ssh-broker-ng/ssh-broker-ng-config-1.dtd
The following sections describe the options available in the Connection Broker configuration file. For more information on the syntax of the configuration file, see Broker Configuration File Syntax.
The SOCKS Proxy also reads three configuration files (if all are available). The files are read in the same order as with the Connection Broker. See the section called “Connection Broker Files” above.
On z/OS, the default configuration file locations are as follows:
the default configuration:
/opt/tectia/etc/ssh-tectia/auxdata/ssh-broker-ng/ssh-socks-proxy-config-default.xml
the global configuration:
/opt/tectia/etc/ssh-socks-proxy-config.xml
the user-specific configuration:
$HOME/.ssh2/ssh-socks-proxy-config.xml
the XML DTD:
/opt/tectia/etc/ssh-tectia/auxdata/ssh-broker-ng/ssh-broker-ng-config-1.dtd
The information in the following sections on the options of the Connection Broker configuration file is also valid for the options of the SOCKS Proxy configuration file. For more information on the syntax of the configuration file, see Broker Configuration File Syntax.
Two kinds of environment variables can be used in the Connection Broker configuration file. In addition to the system-level environment variables, you can use special variables that are Tectia specific. The environment variables take precedence over the special variables. So if an environment variable and a special variable have the same name, the environment variable will be used.
All alphanumeric characters and the underscore '_' sign are allowed in environment variables. The variable name ends to the first character that is not allowed.
You can define for example file or directory paths with environment variables, and they will be expanded to their values as explained below.
Replaced with the value of the environment variable if one has been defined. The
variable is matched case-insensitively. If the variable is not defined, the string
'%VARIABLENAME%
' is the result.
Replaced with the value of the environment variable if one has been defined. The variable is matched case-sensitively on Unix and case-insensitively on Windows. If the variable is not defined, it is replaced with an empty string.
text
Replaced with the value defined for '$VARIABLENAME
' with the
'text
' appended to it.
default_value
}Replaced with the value defined for '$VARIABLENAME
', or
replaced with the 'default_value
' if the variable is not
set.
The Tectia specific special variables are:
Replaced with the currently logged in user name.
Replaced with the currently logged in user name in short format, i.e. without the domain part. Available on Windows.
Replaced with the group name of the currently logged in user.
Replaced with the home directory defined for the currently logged in user.
Replaced with the user identifier defined for the currently logged in user.
Replaced with the group identifier defined for the currently logged in user.
The special variables can also be entered using the Unix format, for example,
$username
.
The Connection Broker configuration file is a valid XML file and starts with the Document Type Declaration.
The root element in the configuration file is secsh-broker
. It can
include general
, default-settings
,
profiles
, static-tunnels
, gui
,
filter-engine,
and logging
elements.
An example of an empty configuration file is shown below:
<!DOCTYPE secsh-broker SYSTEM "ssh-broker-ng-config-1.dtd"> <secsh-broker version="1.0"> <general /> <default-settings /> <profiles /> <static-tunnels /> <gui /> <filter-engine /> <logging /> </secsh-broker>
general
ElementThe general
element contains settings such as the cryptographic library
and the key stores to be used.
The general
element can contain zero or one instance of the following
elements: crypto-lib
, cert-validation
,
key-stores
, user-config-directory
,
protocol-parameters
; and multiple known-hosts
elements.
This element selects the cryptographic library mode to be used. Either the standard
version (standard
) or the FIPS 140-2 certified version
(fips
) of the cryptographic library can be used. The library name is
given as a value of the mode
attribute. By default, standard
cryptographic libraries are used. The OpenSSL cryptographic library is used in the
FIPS mode.
<crypto-lib mode="standard" />
Note | |
---|---|
The FIPS library is not available on z/OS. |
For a list of platforms on which the FIPS library has been validated or tested, see Tectia Client/Server Product Description.
This element defines public-key infrastructure (PKI) settings used for validating
remote server authentication certificates. The element can have the following
attributes: end-point-identity-check
, default-domain
,
http-proxy-url
, socks-server-url
and
max-path-length
.
The end-point-identity-check
attribute specifies whether the client
will verify the server's host name or IP address against the Subject Name or Subject
Alternative Name (DNS Address) specified in the server host certificate. The default
value is yes
. If set to no
, the fields in the server
host certificate are not verified and the certificate is accepted
based on the validity period and CRL check only.
Caution | |
---|---|
Setting |
Alternatively, if set to ask
, the user can decide to either cancel
or continue establishing the connection in case that the server's host name does not
match the one in the certificate.
The default-domain
attribute can be used when the end-point
identity check is enabled. It specifies the default domain part of the remote system
name and it is used if only the base part of the system name is available. The
default-domain
is appended to the system name if it does not contain
a dot (.
).
If the default domain is not specified, the end-point identity check fails, for
example, when a user tries to connect to a host "rock
" giving only the
short host name and the certificate contains the full DNS address
"rock.example.com
".
The http-proxy-url
attribute defines an HTTP proxy and the
socks-server-url
attribute defines a SOCKS server for making LDAP or
OCSP queries for certificate validity.
The address of the server is given as the value of the attribute. The format of the
address is socks://username@socks_server:port/network/netmask,network/netmask
...
(with a SOCKS server) or
http://username@proxy_server:port/network/netmask,network/netmask
...
(with an HTTP proxy).
For example, to make the SOCKS server use host socks.ssh.com
and port 1080
for connections outside of networks
192.196.0.0
(16-bit domain) and 10.100.23.0
(8-bit
domain), and to get these networks connected directly, set
socks-server-url
as follows:
"socks://mylogin@socks.ssh.com:1080/192.196.0.0/16,10.100.23.0/24"
The max-path-length
attribute limits the length of the
certification paths when validating certificates. It can be used to safeguard the paths
or to optimize against the paths getting too long in a deeply hierarchical PKI or when
the PKI is heavily cross-certified with other PKIs. Using the attributes requires
knowing the upper limit of the paths used in certificate validation. For example:
<cert-validation max-path-length="6"> <ldap-server address="ldap://myldap.com" port="389" /> <dod-pki enable="yes" /> <ca-certificate name="CA 1" file="ca-certificate1.crt" /> </cert-validation>
In the example, the path is limited to six certificates, including the end-entity and root CA certificates. If not specified, the default value is 10. Decrease the value to optimize the validation if the maximum length of the encountered paths in the certificate validation is known.
The cert-validation
element can contain multiple
ldap-server
, ocsp-responder
,
crl-prefetch
elements, one dod-pki
element, and
multiple ca-certificate
and key-store
elements. The
elements have to be in the listed order.
This element specifies an LDAP server address
and
port
used for fetching CRLs and/or subordinate CA certificates
based on the issuer name of the certificate being validated. Several LDAP servers
can be specified by using several ldap-server
elements.
CRLs are automatically retrieved from the CRL distribution point defined in the certificate to be verified if the point exists.
The default value for port
is 389
.
This element specifies an OCSP (Online Certificate Status Protocol) responder
service address in URL format with attribute url
. Several OCSP
responders can be specified by using several ocsp-responder
elements.
If the certificate has a valid Authority Info Access extension with an OCSP Responder URL, it will be used instead of this setting. Note that for the OCSP validation to succeed, both the end-entity certificate and the OCSP Responder certificate must be issued by the same CA.
The validity-period
(in seconds) can be optionally defined.
During this time, new OCSP queries for the same certificate are not made but the
old result is used. The default validity period is 0
(a new query
is made every time).
This element instructs Tectia client tools for z/OS to periodically download a CRL from the
specified URL. The url
value can be an LDAP or HTTP URL, or it
can refer to a local file. The file format must be either binary DER or base64,
PEM is not supported.
To download CRLs from the local file system, define the file URL in this format:
file:///absolute/path/name
To download CRLs from an LDAP server, define the LDAP URL in this format:
ldap://ldap.server.com:389/CN=Root%20CA, OU=certification%20authorities,DC=company, DC=com?certificaterevocationlist
Use the interval
attribute to specify how often the CRL is
downloaded. The default is 3600
seconds.
This element defines whether the certificates are required to be compliant
with the US Department of Defense Public-Key Infrastructure (DoD PKI). In
practice, this means that the Digital Signature bit must be set in the Key Usage
of the certificate. The enable
attribute can have a value of
yes
or no
. The default is
no
.
This element defines a certification authority (CA) used in server
authentication. It can have four attributes: name
,
file
, disable-crls
, and
use-expired-crls
.
The name
attribute must contain the name of the CA.
The element must either contain the path to the X.509 CA certificate file as a
value of the file
attribute, or include the certificate as a
base64-encoded ASCII block.
CRL checking can be disabled by setting the disable-crls
attribute to yes
. The default is no
.
Expired CRLs can be used by setting a numeric value (in seconds) for the
use-expired-crls
attribute. The default is 0
(do not use expired CRLs).
This element defines CA certificates stored in an external key store for server authentication. Currently it is used only on z/OS for CA certificates stored in System Authorization Facility (SAF).
The cert-validation/key-store
element has four
attributes: type
, init
,
disable-crls
, and use-expired-crls
.
The type
attribute defines the key store type.
Currently the only supported type is "zos-saf"
.
The init
attribute is the
key-store-provider-specific initialization info.
CRL checking can be disabled by setting the
disable-crls
attribute to yes
. The default is
no
.
Expired CRLs can be used by setting a numeric value (in seconds) for
the use-expired-crls
attribute. The default is 0
(do not use expired CRLs).
For key store configuration examples, see the section called “Key Store Configuration Examples”.
An example of a certificate validation configuration is shown below:
<cert-validation end-point-identity-check="yes" default-domain="example.com" http-proxy-url="http://proxy.example.com:8080"> <ldap-server address="ldap://ldap.example.com:389" /> <ocsp-responder url="http://ocsp.example.com:8090" validity-period="0" /> <crl-prefetch url="file:///full.path.to.crlfile" interval="1800" /> <dod-pki enable="no" /> <ca-certificate name="ssh_ca1" file="ssh_ca1.crt" disable-crls="no" use-expired-crls="100" /> </cert-validation>
This element defines settings for user public-key and certificate authentication.
Under the <general>
element, there can be one
<key-stores>
instance which in turn can have any number of
<key-store>
, <user-keys>
, and
<identification>
elements, and the order of the elements is
free.
Special variables and environment variables can be used when defining the values for the elements. The following variables can be used and they will be expanded as follows:
%U
= %USERNAME%
= user name
%USERNAME-WITHOUT-DOMAIN%
= user name without the domain
part
%IU
= %USERID%
= user ID
%IG
= %GROUPID%
= user group ID
%D
= %HOMEDIR%
= the user's home
directory
%G
= %GROUPNAME%
= the name of the user's
default group
Also environment variables are replaced with their current values. For example it is
possible to use strings $HOME
or %HOME%
to expand to
user's home directory (if environment variable HOME
is set).
Note | |
---|---|
Short alias names (for example, |
Each of the key-store
elements configures one key store
provider. The key-stores/key-store
element can take the following
attributes: type
and init
.
The type
attribute is the key store type. The currently
supported types are "mscapi"
, "pkcs11"
,
"software"
, and "zos-saf"
.
The init
attribute is the initialization info specific to the
key-store-provider. The initialization string can contain special strings
explained above in
key-stores.
For key store configuration examples, see the section called “Key Store Configuration Examples”.
The user-keys
element can be used to override the default
directory for the user keys. The user-keys
element can take the
following attributes:
The directory
attribute defines the directory where the user
private keys are stored. Enter the full path.
The passphrase-timeout
attribute defines the time (in
seconds) after which the passphrase-protected private key will time out, and the
user must enter the passphrase again. The default is 0
, meaning
that the passphrase does not time out. The value of this element should be longer
than the passphrase-idle-timeout
value.
By default, the Connection Broker keeps the passphrase-protected private keys open once
the user has entered the passphrase successfully. This can be changed with the
passphrase timeout options. When passphrase-timeout
is set, the
private key stays open (usable without further passphrase prompts) until the
timeout expires. The passphrase-timeout
attribute sets the hard
timeout, that is set only once when the key is opened and will not be reset even
if the key is used multiple times.
The passphrase-idle-timeout
attribute defines the time (in
seconds) after which the passphrase-protected private key will time out unless the
user accesses or uses the key. The passphrase-idle-timeout
is
reset every time the key is accessed. The default is 0
, meaning
that the passphrase never times out.
Both of the timeout options can be set simultaneously, but notice that if the idle timeout is set longer than the hard timeout, the idle timeout has no effect.
The identification
element can be used to override the
default location of the identification file that defines the user keys. The
identification
element can take the following attributes:
The file
attribute specifies the location of the
identification file. Enter the full path.
The base-path
attribute defines the directory where the
identification file expects the user private keys to be stored. This element can
be used to override the default relative path interpretation of the identification
file (paths relative to the identification file directory).
The passphrase-timeout
attribute defines the time (in
seconds) after which the user must enter the passphrase again. The default is
0
, meaning that the passphrase is not re-requested.
The passphrase-idle-timeout
attribute defines a time (in
seconds) after which the passphrase times out if there are no user actions. The
default is 0
, meaning that the passphrase does not time
out.
The timeout settings affect only those private keys that are listed in the identification file.
Note | |
---|---|
This element is deprecated starting from Tectia client tools for z/OS version 6.1.4. |
This element is supported in configuration for backwards compatibility and used only
if the policy
attribute of the
server-authentication-methods/auth-server-publickey
element under
default-settings
or profiles/profile
is not defined.
In this case, the host key policy is interpreted based on the values of this option and
the host-key-always-ask
and accept-unknown-host-keys
options.
See auth-server-publickey
for details.
Note | |
---|---|
This element is deprecated starting from Tectia client tools for z/OS version 6.1.4. |
This element is supported in configuration for backwards compatibility and used only
if the policy
attribute of the
server-authentication-methods/auth-server-publickey
element under
default-settings
or profiles/profile
is not defined.
In this case, the host key policy is interpreted based on the values of this option and
the strict-host-key-checking
and
accept-unknown-host-keys
options.
See auth-server-publickey
for details.
Note | |
---|---|
This element is deprecated starting from Tectia client tools for z/OS version 6.1.4. |
This element is supported in configuration for backwards compatibility and used only
if the policy
attribute of the
server-authentication-methods/auth-server-publickey
element under
default-settings
or profiles/profile
is not defined.
In this case, the host key policy is interpreted based on the values of this option and
the strict-host-key-checking
and host-key-always-ask
options.
See auth-server-publickey
for details.
Caution | |
---|---|
Consider carefully before enabling this option. Disabling the host-key checks makes you vulnerable to man-in-the-middle attacks. |
This element can be used to change the storage location of the user-specific
configuration files away from the default which is $HOME/.ssh2/
on
Unix, and "%APPDATA%\SSH
" on Windows. It can be used for example,
if you want to store all client-side configurations to a centralized location.
When this element is added to the global configuration file, the Connection Broker reads the following user-specific files in the defined location:
User's key file
User's own configuration files
User's known host keys
User's random_seed file
Windows GUI profile files: 1.ssh2, 2.ssh2
The startup batch file for the sftpg3 client:
ssh_sftp_batch_file
Note | |
---|---|
Stop all existing SSH applications before modifying the
The |
The user-config-directory
option takes an attribute
path
, whose value can be either a directory path or one of the
following variables:
%U
: The user name.
%username%
: The user name.
%username-without-domain%
: The user name without domain
definition.
%D
: The user's home directory.
%homedir%
: The user's home directory.
%USER_CONFIG_DIRECTORY%
: The user-specific configuration
directory.
%IU
: The user's ID, on Unix only
%userid%
: The user's ID, on Unix only
%IG
: The group ID, on Unix only
%groupid%
: The group ID, on Unix only
The default is %USER_CONFIG_DIRECTORY%
. This variable refers to the
user-specific configuration directory: $HOME/.ssh2
on Unix, and
%APPDATA%\SSH
on Windows. The
%USER_CONFIG_DIRECTORY%
variable cannot be used in other
settings.
On Unix, this element can be used to enable checking of file access permissions defined for the global and user-specific configuration files, and for the private keys files. If the permissions are not as expected, the Connection Broker will refuse to start, or to use certain private keys.
By default this setting is disabled. On Windows, this element has no effect.
The file permissions are checked differently, if the
file-access-control
element is set in both the global and user
configuration files, or just in one of them.
In the table: "no"
means file-access-control
enable="no"
. The "-"
sign means that the setting is not
defined in the file at all.
When the file access permissions are checked, the controls are applied as follows:
Expected permissions for the global configuration file: read rights for all, write rights only for the user and group. If the permissions are any wider, the Connection Broker will not start.
Expected permissions for the user configuration file: only the user has read and write rights. If the permissions are any wider, the Connection Broker will not start.
Expected permissions for the private key files: only the user has read and write rights. If the permissions are any wider, keys that do not pass the check will be ignored.
This element contains protocol-specific values that can be used to tune the performance. It should be used only in very specific environments. In normal situations the default values should be used.
The threads
attribute can be used to define the number of threads
the protocol library uses (fast path dispatcher threads). This attribute can be used to
allow more concurrent cryptographic transforms in the protocol on systems with more than
four CPUs. If the value is set to zero, the default value is used.
Example of the threads
attribute:
<protocol-parameters threads="8" />
This element can be used to specify locations for storing the host keys of known
server hosts, and to define the storage format of the host key files. If no
known-hosts
directories are specified, the known host keys are stored
to the default directories.
See the section called “Files”
for the default locations.
On z/OS (only), this element can contain key-store
elements.
This element can be used:
To specify non-default directories that contain the public-key data or public-key files of known server hosts.
To specify a non-default location for OpenSSH-style known_hosts files that contain the public-key data of known server hosts.
(On z/OS) To specify a SAF key store that contains the certificates of known server hosts.
The server host keys are searched in the known-hosts
paths in the
order they are specified in the configuration. The settings of the last defined
known-hosts
element are used when storing new host keys.
If you define any known-hosts
file settings, the default OpenSSH
files will be overridden. So if you wish to make the Connection Broker use both the default OpenSSH
locations and other locations specified in the configuration, you need to specify all
the locations separately.
You can define several known-hosts
elements, and each of them can
contain one or several attributes: path
, directory
,
file
and filename-format
.
The path
attribute requires a full path to the known-hosts file or
directory as the value. For example:
<known-hosts path="/u/username/.ssh/known_hosts" /> <known-hosts path="/etc/ssh2/hostkeys" /> <known-hosts path="/u/username/.ssh2/hostkeys" /> <known-hosts path="/h/username/hostkeys" filename-format="plain" />
The directory
attribute is used to define that known host keys are
saved to a non-default directory. Enter the complete path to the directory as the value.
If the defined directory does not exist, it will be created during the first connection
attempt. If a file is found in its place, the connection will be made but the host key
will not be stored, and the user gets a warning about it. The
filename-format
attribute can be used together with the
directory
setting to define in which format the host key files will
be stored. Example of the directory
attribute:
<known-hosts directory="<path_to_dir>/MyKEYS" filename-format="plain" />
The path
or directory
(whichever is present)
defined in the last known-hosts
element in the configuration file will
be used when storing new known host keys. If both attributes are present in the last
known-hosts
element, the location specified in the
directory
attribute will be used.
The file
attribute is used to point to an OpenSSH-style known_hosts
file. Enter the complete path to the file as the value. If a directory is found in its
place, it is considered an error, and the connection attempt will fail. In case the
known-hosts
element only contains the file
attribute, and the defined OpenSSH known_hosts file exists, the received host keys are
searched first in the defined file, and if not found there, the search continues in the
default Tectia-specific locations.
Example of the file
attribute:
<known-hosts file="<path_to_file>/.ssh2/openSSH_keys" />
An empty file
or path
attribute will disable the
handling of the OpenSSH known_hosts file:
<known-hosts file="" /> or <known-hosts path="" />
The filename-format
attribute defines the format in which new host
key files are stored. The filename-format
attribute is only relevant
for the last specified known-hosts
element and for the default
directory.
The filename-format
attribute takes the values:
hash
(default), plain
, and default
(equals to hash).
With value hash
, the host key files will be stored in format:
keys_<hash>
, for example
"keys_182166d2efe5a134d3fb948646e0b48f780bff6c
".
With value plain
, the file name format will be
key_<port>_<hostname>.pub
, where
<port>
is the port the Secure Shell server is running on and
<hostname>
is the host name you use when connecting to the
server; for example "key_22_my.example.com.pub
".
Setting <known-hosts filename-format="plain" />
changes the
storage format of host key files for the next known-hosts
elements or
for the default storage location if no other known-hosts
elements are
present.
The filename-format="default"
alternative can be used as the last
option when the same known-hosts
element is used to define several
locations for the host keys some of which store the keys in plain format.
For more information on the host key storage formats, see Host Key Storage Formats.
This element defines an external key store for certificates of known server hosts. Currently it is used only on z/OS for server certificates stored in System Authorization Facility (SAF).
The known-hosts/key-store
element has two
attributes: type
and init
.
The type
attribute defines the key store type.
Currently the only supported type is "zos-saf"
.
The init
attribute is the
key-store-provider-specific initialization info.
For key store configuration examples, see the section called “Key Store Configuration Examples”.
This element is reserved for future use.
Example with Software Provider
The software provider handles key pairs stored on disk in standard Secure Shell v2 or legacy OpenSSH formats and X.509 certificates stored in native X.509, PKCS #7, and PKCS #12 formats.
To add a single key file (for example, /u/exa/keys/enigma
and
/etc/my_key
), specify both the private key file and the public key
file:
<key-stores> <key-store type="software" init="key_files(/u/exa/keys/enigma.pub,/u/exa/keys/enigma)" /> <key-store type="software" init="key_files(/etc/my_key.pub,/etc/my_key)" /> </key-stores>
To add all keys from a specific directory (for example all keys from
/u/exa/keys
and /etc/keys
):
<key-stores> <key-store type="software" init="directory(path(/u/exa/keys))" /> <key-store type="software" init="directory(path(/etc/keys))" /> </key-stores>
Example with z/OS SAF Provider
The zos-saf
provider is used for accessing keys stored in the
IBM z/OS System Authorization Facility (SAF).
The initialization string for the zos-saf
provider specifies
the key(s) to be used and it has the following components:
{KEYS([ID(xxx)]RING(xxx) [LABEL(xxx)|DEFAULT])}... [TRUST-ANCHORS]
KEYS(..)
may repeat. The subattributes are:
ID
A SAF user ID signifying the owner of the key ring. If missing, the current user's ID is used.
RING
Key ring name. Mandatory.
LABEL
The SAF key label. If missing, and DEFAULT
is missing, use all
the keys in the key ring.
DEFAULT
Use the key that is marked as the default key on the key ring. Do not specify
together with LABEL
.
TRUST-ANCHORS
Specifies that the certificates in the key ring are trusted CA certificates.
Specifying CA certificates stored in SAF as trusted for server authentication:
<cert-validation> ... <key-store type="zos-saf" init="KEYS(ID(SSHD2) RING(SSH-HOSTCA)) TRUST-ANCHORS" disable-crls="no" use-expired-crls="100" /> </cert-validation>
Specifying server certificates stored in SAF for server authentication:
<known-hosts> ... <key-store type="zos-saf" init="KEYS(ID(USER) RING(SSH-HOSTKEYS))" /> </known-hosts>
Specifying user certificates stored in SAF for user authentication:
<key-stores> <key-store type="zos-saf" init="KEYS(ID(%U) RING(%U))" /> </key-stores>
default-settings
ElementThe default-settings
element defines the default connection-related
settings. Profile-specific settings can override these settings. See the section called “The profiles
Element”.
The default-settings
element can contain zero or one instance of the
following elements in the listed order: ciphers
, macs
,
kexs
, hostkey-algorithms
, rekey
,
authentication-methods
, hostbased-default-domain
,
compression
, proxy
, idle-timeout
,
tcp-connect-timeout
, keepalive-interval
,
exclusive-connection
, server-banners
,
forwards
, extended
, remote-environment
,
server-authentication-methods
,
authentication-success-message
, sftpg3-mode
,
terminal-selection
, terminal-bell
,
close-window-on-disconnect
, quiet-mode
,
checksum
, and address-family
.
The default-settings element can take one attribute:
The user
attribute can be used to define a default user name to be used
when connecting to remote servers. The value of the user
attribute can be one
of the following:
A generic user name that will be used in connections unless another user name is specified in the connection profile settings or in the connection attempt. Note that the user name is treated case sensitively.
"%USERNAME%
" can be used to apply the user name of the currently
logged in user.
In case this option is used but left empty, the Connection Broker will prompt the user for a user name.
The default-settings element can contain the following elements:
This element defines the ciphers that the client will propose to the server. The
ciphers
element can contain multiple cipher
elements.
The ciphers are tried in the order they are specified.
This element selects a cipher name
that the client requests
for data encryption.
The supported ciphers are:
3des-cbc | aes192-ctr | twofish128-cbc |
aes128-cbc | aes256-ctr | twofish192-cbc |
aes192-cbc | arcfour | twofish256-cbc |
aes256-cbc | blowfish-cbc | seed-cbc@ssh.com |
aes128-ctr | twofish-cbc | none (no encryption) |
The default ciphers used by the Connection Broker are, in order:
aes128-cbc
, aes128-ctr,
aes192-cbc
, aes192-ctr,
aes256-cbc
, aes256-ctr,
and
3des-cbc
.
<ciphers> <cipher name="aes128-cbc" /> <cipher name="3des-cbc" /> </ciphers>
This element defines the MACs that the client will propose to the server. The
macs
element can contain multiple mac
elements.
The MACs are tried in the order they are specified.
This element selects a MAC name
that the client requests for
data integrity verification.
The supported MAC algorithms are:
hmac-md5 | hmac-sha224@ssh.com |
hmac-md5-96 | hmac-sha256@ssh.com |
hmac-sha1 | hmac-sha384@ssh.com |
hmac-sha1-96 | hmac-sha2-512 |
hmac-sha2-256 | hmac-sha512@ssh.com |
hmac-sha256-2@ssh.com | none (no data integrity verification) |
The default MACs used by the Connection Broker are, in order:
hmac-sha1 |
hmac-sha1-96 |
hmac-sha2-256 |
hmac-sha256-2@ssh.com |
hmac-sha224@ssh.com |
hmac-sha256@ssh.com |
hmac-sha384@ssh.com |
hmac-sha2-512 |
hmac-sha512@ssh.com |
<macs> <mac name="hmac-sha1" /> </macs>
This element defines the key exchange methods (KEXs) that the client will propose to
the server. The kexs
element can contain multiple kex
elements.
The KEXs are tried in the order they are specified.
This element selects a KEX name
that the client requests for
the key exchange method.
The supported KEX methods are:
diffie-hellman-group1-sha1 |
diffie-hellman-group14-sha1 |
diffie-hellman-group14-sha224@ssh.com |
diffie-hellman-group14-sha256@ssh.com |
diffie-hellman-group15-sha256@ssh.com |
diffie-hellman-group15-sha384@ssh.com |
diffie-hellman-group16-sha384@ssh.com |
diffie-hellman-group16-sha512@ssh.com |
diffie-hellman-group18-sha512@ssh.com |
ecdh-sha2-nistp256 |
ecdh-sha2-nistp384 |
ecdh-sha2-nistp521 |
The default KEX methods used by the Connection Broker are, in order:
diffie-hellman-group1-sha1 |
diffie-hellman-group14-sha1 |
diffie-hellman-group14-sha256@ssh.com |
<kexs> <kex name="diffie-hellman-group1-sha1" /> <kex name="diffie-hellman-group14-sha256@ssh.com" /> </kexs>
This element defines the host key signature algorithms used for server
authentication. The algorithms that will be used are those that are defined in both
Tectia Server and Connection Broker configuration files. This way the use of only certain algorithms, such
as SHA-2, can be enforced by the server. The hostkey-algorithms
element
can contain multiple hostkey-algorithm
elements.
The hostkey algorithms are tried in the order they are specified. Exception: If a host key of a server already exists in the host key store of the client, its algorithm is preferred.
This element selects a host key signature algorithm name
to
be used in server authentication with host keys or certificates.
The supported host key signature algorithms are:
ssh-dss | ecdsa-sha2-nistp521 |
ssh-dss-sha224@ssh.com | x509v3-sign-dss |
ssh-dss-sha256@ssh.com | x509v3-sign-dss-sha224@ssh.com |
ssh-dss-sha384@ssh.com | x509v3-sign-dss-sha256@ssh.com |
ssh-dss-sha512@ssh.com | x509v3-sign-dss-sha384@ssh.com |
ssh-rsa | x509v3-sign-dss-sha512@ssh.com |
ssh-rsa-sha224@ssh.com | x509v3-sign-rsa |
ssh-rsa-sha256@ssh.com | x509v3-sign-rsa-sha224@ssh.com |
ssh-rsa-sha384@ssh.com | x509v3-sign-rsa-sha256@ssh.com |
ssh-rsa-sha512@ssh.com | x509v3-sign-rsa-sha384@ssh.com |
ecdsa-sha2-nistp256 | x509v3-sign-rsa-sha512@ssh.com |
ecdsa-sha2-nistp384 |
The default host key signature algorithms used by the Connection Broker are, in order:
ssh-dss |
ssh-rsa |
ssh-dss-sha256@ssh.com |
ssh-rsa-sha256@ssh.com |
x509v3-sign-dss |
x509v3-sign-rsa |
x509v3-sign-dss-sha256@ssh.com |
x509v3-sign-rsa-sha256@ssh.com |
<hostkey-algorithms> <hostkey-algorithm name="ssh-dss-sha512@ssh.com" /> <hostkey-algorithm name="ssh-rsa-sha224@ssh.com" /> </hostkey-algorithms>
This element specifies the number of transferred bytes
after which
the key exchange is done again. The value "0"
turns rekey requests off.
This does not prevent the server from requesting rekeys, however. The default is
1000000000 (1 GB).
<rekey bytes="1000000000" />
This element specifies the authentication methods that are requested by the
client-side components. The authentication-methods
element can contain
one of each: auth-hostbased
, auth-password
,
auth-publickey
, auth-gssapi
, and
auth-keyboard-interactive
. Alternatively, you can specify multiple
authentication-method
elements. The order of these elements is
free.
The authentication methods are tried in the order the auth-*
or
authentication-method
elements are listed. This means that the least
interactive methods should be placed first.
When several interactive authentication methods are defined as allowed, Tectia client tools for z/OS will alternate between the methods and offers each of them in turn to the server in case the previous method failed.
This element specifies an authentication method name
. It is
included for backwards compatibility. Use the auth-*
elements
instead.
This element specifies that host-based authentication will be used.
The auth-hostbased
element can include a
local-hostname
element.
This element specifies the local host name, as the value of the
name
attribute, that is advertised to the remote server
during host-based authentication.
The remote server can use the client host name as a hint when locating the public key for the client host. This information is not significant to the authentication result, but makes it faster to find the relevant client host key, if the server has such a big storage of host identities, that trying them all would be infeasible.
This element specifies that password authentication will be used.
This element specifies that public-key authentication will be used.
The auth-publickey
element can include a
key-selection
element.
The auth-publickey
element can include a
signature-algorithms
attribute. The attribute defines the
public-key signature algorithms used for client authentication, given as a
comma-separated list. The algorithms that will be used are those that are
defined in both Tectia Server and Connection Broker configuration files. This way the use of only
certain algorithms, such as SHA-2, can be enforced. For a list of the supported
algorithms, see
hostkey-algorithm.
<authentication-methods> <auth-publickey signature-algorithms="ssh-dss,ssh-dss-sha512@ssh.com"/> </authentication-methods>
This element specifies the key selection policy the client uses when
proposing user public keys to the server. The policy
attribute can take the values automatic
(default) and
interactive-shy
.
In the automatic
mode, the client tries keys in the
following order:
Keys with public key available and private key without a passphrase (no user interaction)
Keys with public key available but private key behind a passphrase (one passphrase query)
Keys that need a passphrase to get the public key but private key without passphrase (one user query for each key which is considered and proposed to server, but no user interaction for actual public-key login)
The rest of the keys, that is, keys that need a passphrase to get the public key and also to get the private key
In the interactive-shy
mode, the client does not try
any keys automatically, but it prompts the user to select the key from a
list of available keys. If the authentication with the selected key fails,
the client will prompt the user again, removing the already tried key(s)
from the list. If there is only one key candidate available, the key will be
tried automatically without asking the user.
The key-selection
element can include the
public-key
and issuer-name
elements.
This element can be used to specify that only plain public keys or
only certificates are tried during public-key authentication. The
type
attribute can take the values
plain
and certificate
. The default
is to try both plain public keys and certificates.
This element can be used to filter the user certificates that will
be included in the list presented to the user. The client-side user
certificates can be filtered according to the issuer name that is
compared to the certificate issuers requested or accepted by the
server. The match-server-certificate
attribute takes
values yes
and no
. With value
yes
, Connection Broker tries matching the user certificate
issuer name to the server certificate issuer name. Option
no
means that the issuer names are not used as a
filter. By default, the filtering is not done.
The issuer-name
is useful when a user has several
certificates with different access rights to the same server, for
example for a testing role and for an administrator role. The Connection Broker
chooses the relevant certificates that are applicable on the remote
host, and the user can choose the correct certificate from the
short-listed ones.
This element specifies that keyboard-interactive methods will be used in authentication.
This element specifies that GSSAPI will be used in authentication.
Note | |
---|---|
GSSAPI authentication is not available on z/OS. |
The auth-gssapi
element can take the following
attributes:
The dll-path
attribute specifies where the necessary GSSAPI
libraries are located. If this attribute is not specified, the libraries are
searched for in a number of common locations. The full path to the libraries
should be given, for example,
"/usr/lib/libkrb5.so,/usr/lib/libgssapi_krb5.so"
.
On AIX, the dll-path
should include the archive file, if
applicable, for example,
"<path>/libgssapi_krb5.a(libgssapi_krb5.a.so)"
. The
archive(shared_object)
syntax is not necessary if the library
is a shared object or has been extracted from the shared object.
The allow-ticket-forwarding
attribute defines whether
Tectia client tools for z/OS allows forwarding the Kerberos ticket over several connections. The
attribute can have a value of yes
or no
. The
default is no
.
An example of authentication-methods configuration is shown below:
<authentication-methods> <auth-hostbased> <local-hostname name="host.example.com" /> </auth-hostbased> <auth-gssapi allow-ticket-forwarding="yes"/> <auth-publickey> <key-selection policy="interactive-shy"> <public-key type="plain" /> </key-selection> </auth-publickey> <auth-keyboard-interactive /> <auth-password> <password file="/path/filename" /> </auth-password> </authentication-methods>
This element specifies the host's default domain name (as name
).
This element is used to make sure the fully qualified domain name (FQDN) of the client
host is transmitted to the server when using host-based user authentication.
Note | |
---|---|
This element is not used by Tectia Server 6.4 for IBM z/OS. Instead the |
This element specifies whether the client sends the data compressed (PUT operation). When activated, compression is applied on-the-fly to all data sent out through the connection and on all channels in it.
The name of the compression algorithm and the compression level can be given as
attributes. The name
attribute can be defined as none
(compression not used) or zlib
, currently the only supported algorithm.
By default, compression is not used.
For zlib compression, the level
attribute can be given an integer
from 0
to 9
. The default compression level is
6
, when compression is activated but no level is given (or level is
set to 0
).
Example: to activate maximum level compression of sent data, make the following setting:
<compression name="zlib" level="9" />
Compression can also be activated per connection with command line tools. For information, see the sshg3(1), sftpg3(1) and scpg3(1) man pages.
Note that this compression
setting does not affect received data
(GET operations), but their compression is defined on the Secure Shell server. Tectia Server
always uses compression level 6.
This element defines rules for HTTP proxy or SOCKS servers the client will use for
connections. It has a single attribute: ruleset
.
The format of the attribute value is a sequence of rules delimited by semicolons
(;
). Each rule has a format that resembles the URL format. In a rule,
the connection type is given first. The type can be direct
,
socks
, socks4
, socks5
, or
http-connect
(socks
is a synonym for
socks4
). This is followed by the server address and port. If the port
is not given, the default ports are used: 1080 for SOCKS and 80 for HTTP.
After the address, zero or more conditions delimited by commas (,
)
are given. The conditions can specify IP addresses or DNS names.
direct:///[cond[,cond]...]; socks://server/[cond[,cond]...]; socks4://server/[cond[,cond]...]; socks5://server/[cond[,cond]...]; http-connect://server/[cond[,cond]...]
The IP address/port conditions have an address pattern and an optional port range:
ip_pattern[:port_range]
The ip_pattern
may have one of the following forms:
a single IP address x.x.x.x
an IP address range of the form x.x.x.x-y.y.y.y
an IP sub-network mask of the form x.x.x.x/y
The DNS name conditions consist of a host name which may be a regular expression containing the characters "*" and "?" and a port range:
name_pattern[:port_range]
An example proxy
element is shown below. It causes the server to
access the loopback address and the ssh.com
domain directly, access
*.example
with HTTP CONNECT, and all other destinations with
SOCKS4.
<proxy ruleset="direct:///127.0.0.0/8,*.ssh.com; http-connect://http-proxy.ssh.com:8080/*.example; socks://fw.ssh.com:1080/" />
This element specifies how long idle time (after all connection channels are closed)
is allowed for a connection before automatically closing the connection. The
time
is given in seconds. The type
is always
connection
.
The default setting is 5 seconds. Setting a longer time allows the connection to the server to remain open even after a session (for example, sshg3) is closed. During this time, a new session to the server can be initiated without re-authentication. Setting the time to 0 (zero) terminates the connection immediately when the last channel to the server is closed.
<idle-timeout time="5" />
This element specifies a timeout for the TCP connection. When this setting is made,
connection attempts to a Secure Shell server are stopped after the defined time if the
remote host is down or unreachable. This timeout overrides the default system TCP
timeout, and this timeout setting can be overridden by defining a
tcp-connect-timeout
setting per connection profile (in the
profiles
settings) or per connection (on command line).
The time
is given in seconds. The factory default is 5 seconds.
Value 0 (zero) disables this feature and the default system TCP timeout will be
used.
<tcp-connect-timeout time="5" />
This element specifies an interval for sending keepalive messages to the Secure
Shell server. The time
value is given in seconds. The default setting
is 0, meaning that the keepalive messages are disabled.
<keepalive-interval time="0" />
In the Connection Broker configuration file, the
exclusive-connection
element can be used to specify that a new
connection is opened for each new channel. This setting takes one attribute
enable
, with value yes
or no
. The
default is no
, meaning that open connections are reused for new
channels requested by a client.
On z/OS, the SOCKS Proxy always uses the exclusive connection type, meaning it
opens a new connection for each new channel. This element is not needed in the SOCKS Proxy
configuration file, and irrespective of the value set by the user, the SOCKS Proxy
always interprets this setting as: exclusive-connection enable="yes"
.
This element defines whether the server banner message file (if it exists) is
visible to the user before login. The word yes
or no
is given as the value of the visible
attribute. The default is
yes
.
To eliminate server banners:
<server-banners visible="no" />
This element contains forward
elements that define whether X11 or
agent forwarding (tunneling) are allowed on the client side.
Note | |
---|---|
X11 forwarding is not available on z/OS. |
An example forward configuration, which denies X11 forwarding and allows agent forwarding globally, is shown below:
<forwards> <forward type="x11" state="denied" /> <forward type="agent" state="on" /> </forwards>
This element is reserved for future use.
This element contains environment
elements which define the
environment variables to be passed to the server from the client side. The environment
variables are then set on the server when requesting a command, shell or subsystem.
Note that the server can restrict the setting of environment variables.
This element defines the name and value of the environment variables, and
whether the Connection Broker should process the value. Possible attributes are
name
, value
, and format
.
An example remote environment configuration:
<remote-environment> <environment name="FOO" value="bar" /> <environment name="QUX" value="%Ubaz" format="yes" /> <environment name="ZAPPA" value="%Ubaz" /> </remote-environment>
You can use %U
in the value
to indicate a
user name. When format="yes"
is also defined, the Connection Broker processes
the %U
into the actual user name before sending it to the
server.
Let's assume the user name is joedoe
in this example. The
example configuration results in the following environment variables on the server
side, provided that the server allows setting the environment variables:
FOO=bar QUX=joedoebaz ZAPPA=%Ubaz
You can override the remote environment settings made in the configuration file if
you use the sshg3 command with the following arguments on the
command-line client: --remote-environment
or
--remote-environment-format
For information on the command-line options, see sshg3(1).
This server-authentication-methods
element can be used to force the
Connection Broker to use only certain methods in server authentication. This element can contain
auth-server-publickey
and auth-server-certificate
elements (one of each). The order of these elements is free.
If only auth-server-certificate
is specified, server certificate is
needed. If no server certificate is received, connection fails.
If only auth-server-publickey
is specified, (plain) server public
key is needed. If no server public key is received, connection fails.
If both auth-server-certificate
and
auth-server-publickey
are specified, server certificate is used if
present. Otherwise server public key is used.
The auth-server-certificate
element specifies that
certificates are used for server authentication.
The auth-server-publickey
element specifies that public host
keys are used for server authentication.
Note | |
---|---|
The host key policy settings have changed in version 6.1.4 and are now
defined in the |
The element takes attribute policy
that defines how unknown
server host keys are handled. It can have the following values:
strict
: Connect to the server only if the host key is
found from the host key store and matches.
If the policy is set to strict
, the Connection Broker never adds host
keys to the user's .ssh2/hostkeys
directory upon
connection, and refuses to connect to hosts whose key has changed. This
provides maximum protection against man-in-the-middle attacks. However, it
also means you must always obtain host keys via out-of-band means, which can
be troublesome if you frequently connect to new hosts.
ask
(default): If the server host key is not found from
the host key store, the user will be asked if he wants to accept the host key.
If the host key has changed, the user is warned about it and asked how to
proceed. If the client application is not able to ask the user (for example,
sftpg3 in batch mode, -B
), the
connection will be disconnected.
trust-on-first-use
or tofu
: If the
server host key is not found, it is stored to the user's
.ssh2/hostkeys
directory. If the host key has changed,
the connection will be disconnected.
advisory
: Use of this setting effectively disables server
authentication, which makes the connection vulnerable to active
attackers.
If the server host key is not found in the host key store, it will be
added to the user's .ssh2/hostkeys
directory without user
interaction. If the host key has changed, the connection will be continued
without user interaction. The incident will be audited if logging is
enabled.
When the policy is set to advisory
, the keys from new
hosts are automatically accepted and stored to the host key database without
prompting acceptance from the user. However, changed host keys (from hosts
whose keys are already in the database) are not stored, but they are accepted
for that connection only.
This setting should be used only if logging is enabled for the Connection Broker.
Caution | |
---|---|
Consider carefully before setting the policy to
|
In policy modes other than strict
, if logging is enabled for
the Connection Broker, Tectia client tools for z/OS will log information about changed and new host public
keys with their fingerprints in the syslog (on Unix) or Event Viewer (on
Windows).
Note | |
---|---|
When transparent FTP tunneling or FTP-SFTP conversion is used, accepting the
host key cannot be prompted from the user. Either this setting must be set to
|
If the policy
attribute is not defined, the host key policy
is interpreted based on the values of the old
strict-host-key-checking
, host-key-always-ask
,
and accept-unknown-host-keys
options.
Note | |
---|---|
In version 6.1.4 and later, the host key policy setting in the user-specific configuration file always takes precedence over the setting in the global configuration file. |
The server-authentication-methods/authentication-method
element specifies an authentication method name
. This element is
included for backwards compatibility. Use the auth-server-*
elements instead.
<server-authentication-methods> <authentication-method name="publickey" /> <authentication-method name="certificate" /> </server-authentication-methods>
An example server-authentication-methods
element is shown
below:
<server-authentication-methods> <auth-server-publickey policy="ask" /> <auth-server-certificate /> </server-authentication-methods>
This setting defines whether the AuthenticationSuccessMsg
messages
are output. The authentication-success-message
element takes attribute
enable
with value yes
or no
. The
default is yes
, meaning that the messages are output and logged.
This setting defines how the sftpg3 client behaves when
transferring files. The sftpg3-mode
element takes attribute
compatibility-mode
with the following values:
tectia
(the default) - sftpg3 fransfers
files recursively, meaning that files from the current directory and all its
subdirectories are transferred.
ftp
- the get/put commands are executed as
sget/sput meaning that they transfer a single file; and
commands mget/mput have recursion depth set to 1 meaning that
they only transfer files from the specified directory, not from
subdirectories.
openssh
- commands get/put/mget/mput behave
alike, and the recursion depth is set to 1, meaning that only files from the
specified directory are transferred, not from subdirectories.
The recursion depth can be overridden by using the sftpg3
client's commands get/put/mget/mput with command-line option
--max-depth="LEVEL"
. For more information, see
sftpg3(1).
This element defines how the Tectia terminal behaves when the user selects text with
double-clicks. The element takes one attribute: selection-type
, whose
value can be:
select-words
- double-clicking selects one word at a time, space
and all punctuation characters are used as delimiters. This is the default.
select-paths
- selects strings of characters between spaces,
meaning a selection is extended over characters \/.-_
, so that for example
a path to a file can be selected by double-clicking anywhere in the path.
This element defines whether Tectia terminal repeats audible notifications from the
destination server. This option is only applied to connections with Unix servers. The
element takes one attribute, bell-style
, whose value can be:
none
- no audible notifications are used
pc-speaker
- the user's PC speakers beep when an audible
notification is indicated by the destination server
system-default
- the Tectia terminal sounds the default alerts
defined in the system on the destination server. This is the default.
This element defines that also the Tectia terminal window is to be closed while
disconnecting from a server session by pressing CTRL+D. The element
takes one attribute, enable
, whose value can be yes
or
no
. The default is no
meaning that
CTRL+D closes only the server connection but the Tectia terminal
window remains open.
This setting defines whether the command line clients should suppress warnings,
error messages and authentication success messages. The quiet-mode
element takes attribute enable
with value yes
or
no
. The default is no
, meaning that the errors and
messages are output and logged.
The quiet-mode
element affects command line tools
scpg3, sshg3, and sftpg3.
Enabling the quiet mode here with setting quiet-mode enable="yes"
is
the same as running these clients with option -q
. Note that the
-q
command line parameter will take priority over the
quiet-mode
element set in this configuration file.
The checksum
element can be used to define a default setting for
comparing checksums. This default overwrites the factory setting that checksums are not
checked for files smaller than 32kB.
The checksum
element takes attribute type
, whose
value can be:
yes|YES
- MD5 checksums are checked on files larger than 32kB. This
is the default value.
no|NO
- checksums are not used.
md5|MD5
- only MD5 checksums are checked on files larger than 32kB.
When the --fips
parameter is set with the command line clients
scpg3 and sftpg3, this hash is not used.
sha1|SHA1
- only SHA1 checksums are checked on files larger than
32kB. When the --fips
parameter is set with the command line clients
scpg3 and sftpg3, this hash is used.
md5-force|MD5-FORCE
- MD5 checksums are forced, except when the
--fips
parameter is set with the command line tools
scpg3 and sftpg3.
sha1-force|SHA1-FORCE
- SHA1 checksums are forced on all
files.
checkpoint|CHECKPOINT
- checkpointing is forced on large files that
are transferred one by one.
Note | |
---|---|
If the Connection Broker is started in FIPS mode and the |
Note that checksums can also be defined with the command line clients scpg3 and sftpg3, or with environment variables. The order of priority of the three checksum settings (in case they are different) is as follows, the later one always overwrites the previous value:
checksum
setting in the configuration file
SSH_SFTP_CHECKSUM_MODE
environment variable
Command line arguments
The address-family
element defines the IP address family. Give the
address family as the value of the type
attribute. Tectia client tools for z/OS will
operate using IPv4 (inet
) addressing, IPv6 (inet6
), or
both (any
). The default value for type
is
any
.
profiles
ElementThe profiles
element defines the connection profiles for connecting to
the specified servers. Element profiles
can contain multiple
profile
elements. Each profile defines the connection rules to one server.
The settings in the profile
element override the default connection settings.
When a profile is used for the connection, the settings in the profile override the
default settings. See the section called “The default-settings
Element”.
The profile
element defines a connection profile. It has the
following attributes: id
, name
, host
,
port
, protocol
, host-type
,
connect-on-startup
, user
, and
gateway-profile
.
The profile id
must be a unique identifier that does not change
during the lifetime of the profile.
An additional name
can be given to the profile. This is a free-form
text string. The name can be used for connecting with the profile on the command line,
so define a unique name for each profile.
The host
attribute defines the address of the Secure Shell server
host and it is a mandatory setting. The address can be either an IP address or a domain
name. The value host="*"
can be used to prompt the user to enter the
host address when starting the session.
An empty value host=""
can be used when the
profile is used with transparent TCP or FTP tunneling or FTP-SFTP conversion and the
host name is taken from the application
(filter-engine/rule[@hostname-from-app="yes"]
).
See rule for
details.
The port
is a mandatory setting. It defines the port number of the
Secure Shell server listener. The default port is 22
.
The protocol
is a mandatory setting. It defines the used
communications protocol. Currently the only allowed value is
secsh2
.
If you want to make the connection specified by the profile automatically when the
Connection Broker is started, set the value of the connect-on-startup
attribute to
yes
. In this case, give also the user
attribute (the
user name the connection is made with). You also need to set up some form of
non-interactive authentication for the connection.
The host-type
attribute sets the server type for ASCII (text) file
transfer. This specifies the line break convention that is used for ASCII files. The
default value is default
, meaning that the line break convention is
determined by the local platform. If the client is running on Windows, Windows
compatible line breaks (CR + LF, '\r\n'
) are used. If the client is running
on any other platform, Unix compatible line breaks (LF, '\n'
) are used.
Other possible values for host-type
are windows
(for
Windows remote host) and unix
(for Unix remote host). Define the value
if you are using any other server than Tectia Server.
For FTP-SFTP conversion, set the server type here according to your target server to transfer ASCII files with correct line break convention.
The user
attribute specifies the user name for opening the
connection. The value "%USERNAME%
" can be used to apply the user name
of the currently logged in user. The value user="*"
can be used to
prompt the user to enter the user name when logging in. When the user
attribute is not defined, the user name defined in the default connection settings will
be used.
An empty value user=""
can be used when the
profile is used with transparent FTP tunneling or FTP-SFTP
conversion and the user name is taken from the application
(filter-engine/rule[@username-from-app="yes"]
).
See rule for
details.
The gateway-profile
attribute can be used
to create nested tunnels. The tunnels defined under the local-tunnel
element of the profile, and the tunnels defined under filter-engine
and
static-tunnels
that refer to the profile can be nested. The profile
name through which the connection is made is given as the value of the attribute. The
first tunnel is created using the gateway host profile and from there the second tunnel
is created to the host defined in this profile.
This element gives the path to the remote server host public key file as a
value of the file
attribute.
Alternatively, the public key can be included as a base64-encoded ASCII block.
This element defines the ciphers used with this profile. See ciphers for details.
This element defines the MACs used with this profile. See macs for details.
This element defines the KEXs used with this profile. See kexs for details.
This element defines the hostkey signature algorithms used with this profile.
This element defines the rekeying settings used with this profile. See rekey for details.
This element defines the authentication methods used with this profile. See authentication-methods for details.
This element specifies the identities used in user public-key authentication.
In contrast to the key-stores
element that specifies all the keys
that are available for the Connection Broker, this element can be used to control the keys
that are attempted in authentication when this connection profile is used and to
specify the order in which they are attempted.
The user-identities
element can contain multiple
identity
elements. When multiple identity
elements are used, they are tried out in the order they are listed.
The identity
element has the following attributes:
identity-file
, file
,
hash
, id
, and data
.
The identity-file
attribute specifies that the user
identity is read in the identification file used with public-key
authentication. Enter the full path to the file if it is located somewhere
else than the default identification file directory which is
$HOME/.ssh2
. See also ssh-broker-g3(1).
The file
attribute specifies the path to the public-key
file (primarily) or to a certificate. Enter the full path and file name as
the value.
The hash
attribute is used to enter the hash of the
public key that will be used to identify the related private key. The key
must be available for the Connection Broker The public key hashes of the available keys
can be listed with the ssh-broker-ctl tool. See also
ssh-broker-ctl(1).
The id
attribute is reserved for future use.
The data
attribute is reserved for future use.
An example user-identities
element is shown
below:
<user-identities> <identity identity-file="C:\\ mykey" /> <identity file="$HOME/user/.ssh2/id_dsa_2048_a" /> <identity file="C:\\private_keys\id_dsa_2048_a" /> <identity hash="#a8edd3845005931aaa658b5573609e7d31e23afd" /> </user-identities>
This element defines the compression settings used with this profile. See compression for details.
This element defines the HTTP proxy and SOCKS server settings used with this profile. See proxy for details.
If gateway-profile
has been defined for this profile, the
proxy setting is ignored and the default proxy setting or the proxy setting of the
gateway profile is used instead.
This element defines the idle timeout settings used with this profile. See idle-timeout for details.
This element defines the TCP connection timeout for this profile. The timeout is used to terminate connection attempts to Secure Shell servers that are down or unreachable. The default value is 5 seconds. See tcp-connect-timeout for details.
This element defines an interval for sending keepalive messages to the Secure Shell server. The setting applies to this profile. The default value is 0, meaning that no keepalive messages are sent. See keepalive-interval for details.
In the Connection Broker configuration, this
element defines whether a new connection is
opened for each new channel when a connection is made with this profile. This
setting takes one attribute enable
, with value
yes
or no
. The default is no
,
meaning that open connections are reused for new channels requested by a client.
See also
exclusive-connection.
On z/OS, the SOCKS Proxy always uses the exclusive connection type, meaning
it opens a new connection for each new channel. This element is not needed in the
SOCKS Proxy configuration file, and irrespective of the value set by the user, the SOCKS Proxy
always interprets this setting as:
exclusive-connection enable="yes"
.
This element defines the server banner setting used with this profile. See server-banners for details.
This element defines the forwards allowed with this profile. See forwards for details.
The tunnels
element defines the tunnels that are opened when
a connection with this profile is made. The element can contain multiple
local-tunnel
and remote-tunnel
elements.
This element defines a local tunnel (port forwarding) that is opened
automatically when a connection is made with the connection profile. It has
five attributes: type
, listen-port
,
listen-address
, dst-host
,
dst-port
, and allow-relay
.
The type
attribute defines the type of the tunnel. This
can be tcp
(default, no special processing),
ftp
(temporary forwarding is created for FTP data
channels, effectively securing the whole FTP session), or
socks
(Tectia client tools for z/OS will act as a SOCKS server for other
applications, creating forwards as requested by the SOCKS
transaction).
The listen-port
attribute defines the listener port
number on the local client.
The listen-address
attribute can be used to define
which network interfaces on the client should be listened. Its value can be
an IP address belonging to an interface on the local host. Value
0.0.0.0
listens to all interfaces. The default is
127.0.0.1
(localhost loopback address on the client).
Setting any other value requires setting
allow-relay="yes"
.
For address-family
option inet6
, the
default listen address is ::1
. To listen on all interfaces,
specify ::
. For address-family
option
any
, the listen address is both
127.0.0.1
and ::1
by default; to listen
on all interfaces, specify ::
.
Whenever a connection is made to the specified listener, the connection
is tunneled over Secure Shell to the remote server and another connection is
made from the server to a specified destination host and port
(dst-host
, dst-port
). The connection
from the server onwards will not be secure, it is a normal TCP
connection.
The dst-host
and dst-port
attributes
define the destination host address and port. The value of
dst-host
can be either an IP address or a domain name.
The default is 127.0.0.1
(localhost = server host).
The allow-relay
attribute defines whether connections
to the listened port are allowed from outside the client host. The default
is no
. If you use allow-relay="yes"
, it
will check also the listen-address
setting.
This element defines a remote tunnel (port forwarding) that is opened
automatically when a connection is made with the connection profile. It has
four attributes: type
, listen-port
,
listen-address
, dst-host
,
dst-port
, and allow-relay
.
The type
attribute defines the type of the tunnel. This
can be either tcp
(default, no special processing) or
ftp
(temporary forwarding is created for FTP data
channels, effectively securing the FTP session between the client and
server).
The listen-port
attribute defines the listener port
number on the remote server.
The listen-address
attribute can be used to define
which network interfaces on the server should be listened. Its value can be
an IP address belonging to an interface on the server host. Value
0.0.0.0
listens to all interfaces. The default is
127.0.0.1
(localhost loopback address on the server).
Setting any other value requires that
allow-relay="yes"
.
For address-family
option inet6
, the
default listen address is ::1
. To listen on all interfaces,
specify ::
. For address-family
option
any
, the listen address is both
127.0.0.1
and ::1
by default; to listen
on all interfaces, specify ::
.
Whenever a connection is made to this listener, the connection is
tunneled over Secure Shell to the local client and another connection is
made from the client to a specified destination host and port
(dst-host
, dst-port
). The connection
from the client onwards will not be secure, it is a normal TCP
connection.
The dst-host
and dst-port
attributes
define the destination host address and port. The value of
dst-host
can be either an IP address or a domain name.
The default is 127.0.0.1
(localhost = client host).
The allow-relay
attribute defines whether connections
to the listener port are allowed from outside the server host. The default
is no
.
This element is reserved for future use.
This element defines the remote environment settings used with this profile.
Within the remote-environment
element, define an
environment
element for each environment variable to be passed
to the server.
See remote-environment
for details.
This element defines the server authentication methods allowed with this profile. See server-authentication-methods for details.
This element can be used to specify a user password that the client will send as a response to password authentication.
The password can be given directly in the string
attribute,
or a path to a file containing the password can be given in the
file
attribute, or a path to a program or a script that outputs
the password can be given in the command
attribute.
When using the command
option to refer to a shell script,
make sure the script also defines the user's shell, and outputs the actual
password. Otherwise the executed program fails, because it does not know what
shell to use for the shell script. For example, if the password string is defined
in a file named my_password.txt
, and you want to use the bash
shell, include these lines in the script:
#!/usr/bash cat /full/pathname/to/my_password.txt
Caution | |
---|---|
If the password is given using this option, it is extremely important that
the |
Note | |
---|---|
Any password given with the command-line options will override this setting. |
An example connection profile is shown below:
<profile name="rock" id="id1" host="rock.example.com" port="22" connect-on-startup="no" user="doct"> <hostkey file="key_22_rock.pub"> </hostkey> <authentication-methods> <auth-publickey /> <auth-password /> </authentication-methods> <server-authentication-methods> <auth-server-publickey policy="strict" /> </server-authentication-methods> <server-banners visible="yes" /> <forwards> <forward type="agent" state="on" /> <forward type="x11" state="on" /> </forwards> <tunnels> <local-tunnel type="tcp" listen-port="143" dst-host="imap.example.com" dst-port="143" allow-relay="no" /> </tunnels> <remote-environment> <environment name="FOO" value="bar" /> <environment name="QUX" value="%Ubaz" format="yes" /> <environment name="ZAPPA" value="%Ubaz" /> </remote-environment> </profile>
static-tunnels
ElementThe static-tunnels
setting is used to configure the behavior of the
automatic tunnels. You can create listeners for local tunnels automatically when the Connection Broker
starts up. The actual tunnel is formed the first time a connection is made to the listener
port. If the connection to the server is not open at that time, it will be opened
automatically as well.
The static-tunnels
element can contain any number of
tunnel
elements.
The tunnel
element specifies a static tunnel. It has the following
attributes: type
, listen-port
,
listen-address
, dst-host
, dst-port
,
allow-relay
, and profile
.
The type
attribute defines the type of the tunnel. This
can be either tcp
, ftp
, or
socks-proxy
.
tcp
specifies a listener for generic TCP tunneling
ftp
specifies a listener for FTP tunneling (also the FTP data
channels are tunneled)
socks-proxy
specifies a listener that acts as a SOCKS proxy
towards the client applications. The traffic coming to the proxy is filtered using
filter rules. When this option is used, a filter-engine
element
must be defined.
See filter-engine.
The listen-port
attribute defines the listener port number on the
local client.
The listen-address
attribute can be used to define which network
interfaces on the client should be listened. Its value can be an IP address belonging to
an interface on the local host. Value 0.0.0.0
listens to all
interfaces. The default is 127.0.0.1
(localhost loopback address on the
client). Setting any other value requires that
allow-relay="yes"
.
For address-family
option inet6
, the default
listen address is ::1
. To listen on all interfaces, specify
::
. For address-family
option any
,
the listen address is both 127.0.0.1
and ::1
by
default; to listen on all interfaces, specify ::
.
The dst-host
and dst-port
attributes define the
destination host address and port. The value of dst-host
can be either
an IP address or a domain name. The default is 127.0.0.1
(localhost =
server host).
The allow-relay
attribute defines whether connections to the
listened port are allowed from outside the client host. The default is
no
.
The profile
attribute specifies the connection profile ID that is
used for the tunnel.
<static-tunnels> <tunnel type="tcp" listen-address="127.0.0.1" listen-port="9000" dst-host="st.example.com" dst-port="9000" allow-relay="no" profile="id1" /> </static-tunnels>
filter-engine
ElementThe filter-engine
element defines the filter rules for FTP-SFTP
conversion and transparent FTP tunneling. On z/OS, also the static-tunnels
element with a socks-proxy
tunnel needs to be defined for the transparent FTP
security to work.
Note | |
---|---|
The SOCKS Proxy reads the |
The top level element is filter-engine
. It has three attributes:
ip-generate-start
, ip6-generate-start
and
ftp-filter-at-signs
(used with Tectia ConnectSecure, only).
The ip-generate-start
attribute defines the start address of the pseudo
IPv4 address space. Pseudo IPs are generated by the Connection Broker when applications do the DNS query
through the SSH connection capture component.
The ip6-generate-start
attribute is similar to
ip-generate-start
, but it defines the start address of the pseudo IPv6
address space.
Note | |
---|---|
Under the |
The network
element specifies a "location" where Tectia client tools for z/OS is
running. By using the network
element, you can implement
location-awareness for Tectia client tools for z/OS. It has five attributes: id
,
address
, domain
, ip-generate-start
and ip6-generate-start
.
The id
attribute specifies a unique identifier for the
network
element. The address
attribute specifies the
address of the network. It can be missing or empty, in which case it is not used. The
domain
attribute contains the domain name of the computer. It can
also be missing or empty, in which case it is not used. The
ip-generate-start
attribute defines the start address of the pseudo
IPv4 space. If it is defined here, it overrides the ip-generate-start
attribute of the filter-engine
element. The
ip6-generate-start
attribute is similar to
ip-generate-start
, but it defines the start address of the pseudo
IPv6 address space.
Note | |
---|---|
The |
The dns
element creates a DNS rule for the filter engine. It has
six attributes: id
, network-id
,
application
, host
, ip-address
, and
pseudo-ip
.
For their descriptions, see rule below.
Note | |
---|---|
The |
The filter
element specifies an action for a connection. It has the
following attributes: dns-id
, ports
,
action
, profile-id
, destination
,
destination-port
, fallback-to-plain
.
The dns-id
attribute is a reference to a dns
element.
For the descriptions of the other attributes, see rule below.
The rule
element specifies how a filtered connection will be
handled. It has the following attributes: application
,
host
, ip-address
, pseudo-ip
,
ports
, action
, profile-id
,
destination
, destination-port
,
username
, hostname-from-app
,
username-from-app
, fallback-to-plain
,
show-sftp-server-banner
.
The application
attribute can be used to specify one or more
applications to which the rule is applied. This can be a regular expression using the
egrep syntax.
For information on the syntax, see
Appendix C.
The host
attribute specifies a target host name. It can be a
regular expression using the egrep syntax.
The ip-address
attribute specifies the target host IP address. It
can be a regular expression using the egrep syntax. In this case the Connection Broker does the
string matching with the assumption that the IP address is written in its canonical
form. If both the host name and the IP address are defined, the host
attribute takes precedence and the ip-address
attribute is
ignored.
The pseudo-ip
setting has the following effects when the
ip-address
is left empty and the host
matches:
When pseudo-ip="yes"
, the Connection Broker assigns a pseudo IP address for
the target host and Tectia Server resolves the real IP address. The pseudo IP addresses
should be used when accessing an internal network from the outside, because name
resolution for the machines in the internal network is not available from the
outside.
When pseudo-ip="no"
, a normal DNS query is made for the target
host name. The default value is no
.
The ports
attribute can be a single port or a range. A range is
specified with a hyphen between two integers (for example
"21-25"
).
The action
attribute specifies the action to be done when a filter
matches. Its value can be DIRECT
, BLOCK
,
TUNNEL
, FTP-TUNNEL
, or
FTP-PROXY
.
DIRECT
causes the connection to be made directly as plaintext
without tunneling or FTP-SFTP conversion.
BLOCK
causes the connection to be blocked.
FTP-TUNNEL
activates transparent FTP tunneling
FTP-PROXY
causes the FTP-SFTP conversion to start and a
connection to be made to the Secure Shell SFTP server.
The profile-id
attribute can be used to specify the connection
profile that defines the connection settings.
If the profile-id
attribute is left empty and
hostname-from-app="yes"
is specified, the Secure Shell connection is
made to the server specified by the client application using default settings. If a
profile-id
is specified and also
hostname-from-app="yes"
is specified, or the referred profile has
*
(an asterisk) or empty as the value of the host
attribute, the Secure Shell connection is made to the server specified by the client
application using the profile settings.
The destination
and destination-port
attributes
can be used to define a static destination address and port number that will be used as
the end point of the connection instead of the original address and port given by the
application.
The username
attribute can be used to define the user name used for
connecting to the Secure Shell server, or you can define the path from where the Connection Broker
should retrieve the user name.
The hostname-from-app
attribute defines whether the Connection Broker should
extract the Secure Shell server's host name from data sent by the application, or use a
Secure Shell server defined by the connection profile in profile-id
.
The value is yes
or no
, and the default is
no
.
When hostname-from-app="no"
, the tunnel
or FTP-SFTP conversion will be created to the
Secure Shell server specified in the connection profile referred in the
profile-id
attribute. Note that with transparent tunneling, the
connection from the Secure Shell server to the final destination application will be
unsecured and in plaintext. To achieve end-to-end security, the Secure Shell server
should reside on the same host as the application.
When hostname-from-app="yes"
, the tunnel
or FTP-SFTP conversion will be created to the
destination server specified by the application. This setting can be used with both FTP
and TCP tunneling and FTP-SFTP conversion. When using
hostname-from-app="yes"
, it is no longer necessary to create a
separate connection profile for each destination host. Note that this requires that a
Secure Shell server is installed to each destination server
(or, in the case of tunneling, that
fallback-to-plain
is enabled to allow direct connections to those
servers that do not have Secure Shell installed).
The username-from-app
attribute defines whether the FTP tunneling
or FTP-SFTP conversion extracts the user name from data sent by the FTP application. The
value is yes
or no
. The default is
no
.
When username-from-app="yes"
, the user name received from the FTP
client application is used. This setting can be used with FTP tunneling and FTP-SFTP
conversion. This setting will override any user name settings made in a related
connection profile. When username-from-app="no"
, the user name is taken
from the connection profile defined with the profile-id
attribute.
In transparent tunneling, the
fallback-to-plain
attribute can be used to define whether a direct
(unsecured) connection is used if creating the tunnel fails. The default value is
no
. Normally, when
creating the secure tunnel
fails when applying a filter rule, the Connection Broker will return an error about not being able to
establish a connection.
Note | |
---|---|
Do not enable the |
In FTP-SFTP conversion, if the target SFTP server is configured to send a banner to
the client, the show-sftp-server-banner
attribute can be used to make
the Connection Broker forward the SFTP server banner to the FTP client. The allowed values for
show-sftp-server-banner
are yes
and
no
. The default value is no
(the SFTP server banner
is not forwarded to the FTP client).
Note | |
---|---|
If you have made changes to the Connection Broker default configuration, make sure that
showing the server banner is enabled ( |
Note | |
---|---|
Sending the SFTP server banner to the FTP client will cause an extra connection opening to the target SFTP server for retrieving the banner message. |
In case of a failure in retrieving the banner message from the target SFTP server, the banner that Connection Broker forwards to the FTP client includes an error description, a default banner, and the following text:
Can't fetch banner from SFTP Server
logging
ElementThe logging
element changes the logging settings that define the log
event severities and logging facilities. The element contains one or more
log-target
and log-events
elements.
This element specifies the log target for auditing. By default, the broker does not log anything. This element can be used to direct log data to a file or syslog.
The log-target
element can have file
and
type
as attributes.
The type
attribute specifies the logging facility where the audit
data is output to. The value can be file
, syslog
or
discard
.
The file
attribute sets the file system path where the audit data
is written to. If the type
attribute has syslog
or
discard
set, the file
attribute is not
allowed.
This element sets the severity and facility of different logging events. The events have reasonable default values, which are used if no explicit logging settings are made. This setting allows customizing the default values.
The element can also contain one or more log-target
elements. When
defined here, the log-target
element will override the definition given
in the logging
/log-target
.
For the events, facility
and severity
can be set
as attributes. The events itself should be listed inside the log-events
element.
The facility can be normal
, daemon
,
user
, auth
, local0
,
local1
, local2
, local3
,
local4
, local5
, local6
,
local7
, or discard
. Setting the facility to
discard
causes the server to ignore the specified log events.
The severity can be informational
, notice
,
warning
, error
, critical
,
security-success
, or security-failure
.
Any events that are not specifically defined in the configuration file will use the
default values. The defaults can be overridden for all remaining events by giving an
empty log-events
element after all other definitions and by setting a
severity value for it.
In the names of log events, the characters '*' and '?' can be used as wildcards.
An example logging configuration that logs all events, which are programmed to be logged
by default, both to /tmp/foo
and to syslog
.
<logging> <log-target file="/tmp/foo" /> <log-target type="syslog" /> </logging>
An example logging configuration in which events are logged to
/tmp/foo
, except those whose event name matches
"Key_store_*
", which will be discarded.
<logging> <log-target file="/tmp/foo" /> <log-events facility="discard"> Key_store_* </log-events> </logging>