openldap/doc/man/man5/lloadd.conf.5

.TH LLOADD.CONF 5 "RELEASEDATE" "OpenLDAP LDVERSION"
.\" Copyright 1998-2022 The OpenLDAP Foundation All Rights Reserved.
.\" Copying restrictions apply.  See COPYRIGHT/LICENSE.
.\" $OpenLDAP$
.SH NAME
lloadd.conf \- configuration file for lloadd, the stand-alone LDAP daemon
.SH SYNOPSIS
ETCDIR/lloadd.conf
.SH DESCRIPTION
The file
.B ETCDIR/lloadd.conf
contains configuration information for the
.BR lloadd (8) daemon.
.LP
The
.B lloadd.conf
file consists of a series of global configuration options that apply to
.B lloadd
as a whole (including all backends), followed by zero or more
backend definitions that contain information specific how a backend
instance should be contacted.
The configuration options are case-insensitive;
their value, on a case by case basis, may be case-sensitive.
.LP
The general format of
.B lloadd.conf
is as follows:
.LP
.nf
    # comment - these options apply to the server as a whole
    <global configuration options>
    # first backend definition
    backend-server <backend 1 definition>
    # subsequent backend definitions
    ...
.fi
.LP
As many backend servers may be configured as desired.
.LP
If a line begins with white space, it is considered a continuation
of the previous line.  No physical line should be over 2000 bytes
long.
.LP
Blank lines and comment lines beginning with
a `#' character are ignored.  Note: continuation lines are unwrapped
before comment processing is applied.
.LP
Arguments on configuration lines are separated by white space. If an
argument contains white space, the argument should be enclosed in
double quotes.  If an argument contains a double quote (`"') or a
backslash character (`\\'), the character should be preceded by a
backslash character.
.LP
The specific configuration options available are discussed below in the
Global Configuration Options and General Backend Options.
Refer to the "OpenLDAP Administrator's Guide" for more
details on the lloadd configuration file.

.SH SLAPD INTEGRATION
Note that when
.B lloadd
is configured as a
.B slapd
module, any option that shares the same name as an option in
.BR slapd.conf (5),
the
.B slapd
interpretation wins and the
.B lloadd
option mentioned is unavailable through
.BR slapd.conf (5)
directly, instead, it would have to be configured via a dedicated attribute in
cn=config. In particular, unless the
.B TLSShareSlapdCTX
option is set,
.B lloadd
keeps its own TLS context which cannot be configured except
through the dynamic configuration.

An additional option is available when running as a
.B slapd
module:
.TP
.B listen "<listen URIs>"
The URIs the Load Balancer module should listen on. Must not overlap with the
ones that
.B slapd
uses for its own listening sockets. The related
.B cn=config
attribute is
.B olcBkLloadListen
with each URI provided as a separate value. No changes to this attribute made
after the server has started up will take effect until it is restarted.

.SH GLOBAL CONFIGURATION OPTIONS
Options described in this section apply to all backends. Arguments that should
be replaced by actual text are shown in brackets <>.
.TP
.B argsfile <filename>
The (absolute) name of a file that will hold the
.B lloadd
server's command line (program name and options).
.TP
.B concurrency <integer>
Specify a desired level of concurrency.  Provided to the underlying
thread system as a hint.  The default is not to provide any hint.
.\" .TP
.\" .B gentlehup { on | off }
.\" A SIGHUP signal will only cause a 'gentle' shutdown-attempt:
.\" .B Lloadd
.\" will stop listening for new connections, but will not close the
.\" connections to the current clients.  Future write operations return
.\" unwilling-to-perform, though.  Lloadd terminates when all clients
.\" have closed their connections (if they ever do), or - as before -
.\" if it receives a SIGTERM signal.  This can be useful if you wish to
.\" terminate the server and start a new
.\" .B lloadd
.\" server
.\" .B with another database,
.\" without disrupting the currently active clients.
.\" The default is off.  You may wish to use
.\" .B idletimeout
.\" along with this option.
.\" .TP
.\" .B idletimeout <integer>
.\" Specify the number of seconds to wait before forcibly closing
.\" an idle client connection.  A idletimeout of 0 disables this
.\" feature.  The default is 0. You may also want to set the
.\" .B iotimeout
.\" option.
.TP
.B feature <feature> [...]
Switch additional features supported by the LDAP Load Balancer on.
Supported features are:
.RS
.RS
.PD 0
.TP
.B proxyauthz
when proxying an operation, pass the client's authorized identity using
the proxy authorization control (RFC 4370). No control is added to the
operation if initiated by a client whose bound identity matches the identity
configured in
.B bindconf
(no normalisation of the DN is attempted).

If SASL binds are issued by clients and this feature is enabled, backend
servers need to support LDAP Who Am I? extended operation for the Load Balancer
to detect the correct authorization identity.
.\" .TP
.\" .B vc
.\" when receiving a bind operation from a client, pass it onto a backend
.\" as a verify credentials external operation request. With this enabled,
.\" the
.\" .BR backend 's
.\" .B bindconns
.\" option has no effect as there is no need to maintain dedicated bind
.\" connections anymore.
.PD
.RE
.RE
.TP
.B include <filename>
Read additional configuration information from the given file before
continuing with the next line of the current file.
.TP
.B io-threads <integer>
Specify the number of threads to use for the connection manager.
The default is 1 and this is typically adequate for up to 16 CPU cores.
The value should be set to a power of 2.

If modified after server starts up, a change to this option will not take
effect until the server has been restarted.
.TP
.B logfile <filename>
Specify a file for recording lloadd debug messages. By default these messages
only go to stderr, are not recorded anywhere else, and are unrelated to
messages exposed by the
.TP
.B logfile-format debug | syslog-utc | syslog-localtime
Specify the prefix format for messages written to the logfile. The debug
format is the normal format used for slapd debug messages, with a timestamp
in hexadecimal, followed by a thread ID.  The other options are to
use syslog(3) style prefixes, with timestamps either in UTC or in the
local timezone. The default is debug format.
.B loglevel
configuration parameter. Specifying a logfile copies messages to both stderr
and the logfile.
.TP
.B logfile-only on | off
Specify that debug messages should only go to the configured logfile, and
not to stderr.
.TP
.B logfile-rotate <max> <Mbytes> <hours>
Specify automatic rotation for the configured logfile as the maximum
number of old logfiles to retain, a maximum size in megabytes to allow a
logfile to grow before rotation, and a maximum age in hours for a logfile
to be used before rotation. The maximum number must be in the range 1-99.
Setting Mbytes or hours to zero disables the size or age check, respectively.
At least one of Mbytes or hours must be non-zero. By default no automatic
rotation will be performed.
.TP
.B loglevel <integer> [...]
Specify the level at which debugging statements and operation
statistics should be syslogged (currently logged to the
.BR syslogd (8)
LOG_LOCAL4 facility).
They must be considered subsystems rather than increasingly verbose
log levels.
Some messages with higher priority are logged regardless
of the configured loglevel as soon as any logging is configured.
Log levels are additive, and available levels are:
.RS
.RS
.PD 0
.TP
.B 1
.B (0x1 trace)
trace function calls
.TP
.B 2
.B (0x2 packets)
debug packet handling
.TP
.B 4
.B (0x4 args)
heavy trace debugging (function args)
.TP
.B 8
.B (0x8 conns)
connection management
.TP
.B 16
.B (0x10 BER)
print out packets sent and received
.\" .TP
.\" .B 32
.\" .B (0x20 filter)
.\" search filter processing
.TP
.B 64
.B (0x40 config)
configuration file processing
.\" .TP
.\" .B 128
.\" .B (0x80 ACL)
.\" access control list processing
.TP
.B 256
.B (0x100 stats)
connections, LDAP operations, results (recommended)
.TP
.B 512
.B (0x200 stats2)
stats log entries sent
.\" .TP
.\" .B 1024
.\" .B (0x400 shell)
.\" print communication with shell backends
.\" .TP
.\" .B 2048
.\" .B (0x800 parse)
.\" entry parsing
\".TP
\".B 4096
\".B (0x1000 cache)
\"caching (unused)
\".TP
\".B 8192
\".B (0x2000 index)
\"data indexing (unused)
.\" .TP
.\" .B 16384
.\" .B (0x4000 sync)
.\" LDAPSync replication
.TP
.B 32768
.B (0x8000 none)
only messages that get logged whatever log level is set
.PD
.RE
The desired log level can be input as a single integer that combines
the (ORed) desired levels, both in decimal or in hexadecimal notation,
as a list of integers (that are ORed internally),
or as a list of the names that are shown between parentheses, such that
.LP
.nf
    loglevel 513
    loglevel 0x201
    loglevel 512 1
    loglevel 0x200 0x1
    loglevel stats trace
.fi
.LP
are equivalent.
The keyword
.B any
can be used as a shortcut to enable logging at all levels (equivalent to \-1).
The keyword
.BR none ,
or the equivalent integer representation, causes those messages
that are logged regardless of the configured loglevel to be logged.
In fact, if loglevel is set to 0, no logging occurs,
so at least the
.B none
level is required to have high priority messages logged.

The loglevel defaults to \fBstats\fP.
This level should usually also be included when using other loglevels, to
help analyze the logs.
.RE
.TP
.B pidfile <filename>
The (absolute) name of a file that will hold the
.B lloadd
server's process ID (see
.BR getpid (2)).
.TP
.B sockbuf_max_incoming_client <integer>
Specify the maximum LDAP PDU size accepted coming from clients.
The default is 262143.
.TP
.B sockbuf_max_incoming_upstream <integer>
Specify the maximum LDAP PDU size accepted coming from upstream
connections.
The default is 4194303.
.TP
.B tcp-buffer [listener=<URL>] [{read|write}=]<size>
Specify the size of the TCP buffer.
A global value for both read and write TCP buffers related to any listener
is defined, unless the listener is explicitly specified,
or either the read or write qualifiers are used.
See
.BR tcp (7)
for details.
Note that some OS-es implement automatic TCP buffer tuning.
.TP
.B threads <integer>
Specify the maximum size of the primary thread pool.
The default is 16; the minimum value is 2.
.TP
.B threadqueues <integer>
Specify the number of work queues to use for the primary thread pool.
The default is 1 and this is typically adequate for up to 8 CPU cores.
The value should not exceed the number of CPUs in the system.
.TP
.B max_pdus_per_cycle <integer>
If set to 0, PDUs are handled by the I/O threads directly, otherwise
a task is queued to be picked up by the thread pool. This task will
process PDUs from the connection until there is no more data to be
read or this limit is reached when the I/O thread can pick it up again.
Very high values have a potential to cause some connections to be
starved in a very high-bandwidth environment. The default is 1000.
.TP
.B client_max_pending <integer>
Will cause the load balancer to limit the number unfinished operations for each
client connection. The default is 0, unlimited.
.TP
.B iotimeout <integer>
Specify the number of milliseconds to wait before forcibly closing
a connection with an outstanding write. This allows faster recovery from
various network hang conditions.  An iotimeout of 0 disables this feature.
The default is 10000.
.TP
.B write_coherence <integer>
Specify the number of seconds after a write operation is finished that
.B lloadd
will direct operations exclusively to the last selected backend. A write
operation is anything not handled internally (certain exops, abandon),
except search, compare and bind operations. Bind operations also reset this
restriction. The default is 0, write operations do not restrict selection. When
negative, the restriction is not time limited and will persist until the next
bind.
.TP
.B restrict_exop <OID> <action>
Tell
.B lloadd
that extended operation with a given OID should be handled in a specific way.
OID
.B 1.1
is special, setting a default (only for operations not handled internally).
The meaning of the
.B <action>
argument is the same as in
.B restrict_control
below.
.TP
.B restrict_control <OID> <action>
Tell
.B lloadd
that a control with a given OID attached to any operation should be handled in
a specific way according to the
.B <action>
argument. At the moment, only operations passed intact are inspected in
this way, in particular, controls on bind and extended operations are
.B not
checked.

In order of descending priority (the control with highest priority action
wins), this is the action made:
.RS
.RS
.PD 0
.TP
.B reject
operations that carry this control will be rejected.
.TP
.B connection
once an upstream is selected, every future operation from this client will be
directed to the same connection. Useful when state is shared between client and
upstream that the load balancer doesn't track.
.TP
.B backend
like
.B write
except this does not time out.
.TP
.B write
this is treated like a write operation (see
.BR write_coherence )
above.
.TP
.B ignore
does not influence restrictions, useful when changing the global exop default.
This is the default handling for exops/controls not handled by the load balancer
internally.
.PD
.RE

.SH TLS OPTIONS
If
.B lloadd
is built with support for Transport Layer Security, there are more options
you can specify.

.TP
.B TLSShareSlapdCTX { on | off }
If set to no (the default),
.B lloadd
will use its own TLS context (needs to be configured via
.B cn=config
unless
.B lloadd
is run as a standalone daemon). If enabled, the options for
.B slapd
apply instead, since the
.BR slapd 's
TLS context is used then.

.LP

The following options are available only when compiled as a standalone daemon.
When compiled as a
.BR slapd (8)
module, the cn=config equivalents need to be used if a separate TLS context for
the module is needed, otherwise use the
.B TLSShareSlapdCTX
option.

.TP
.B TLSCipherSuite <cipher-suite-spec>
Permits configuring what ciphers will be accepted and the preference order.
<cipher-suite-spec> should be a cipher specification for the TLS library
in use (OpenSSL, GnuTLS, or Mozilla NSS).
Example:
.RS
.RS
.TP
.I OpenSSL:
TLSCipherSuite HIGH:MEDIUM:+SSLv2
.TP
.I GnuTLS:
TLSCiphersuite SECURE256:!AES-128-CBC
.RE

To check what ciphers a given spec selects in OpenSSL, use:

.nf
	openssl ciphers \-v <cipher-suite-spec>
.fi

With GnuTLS the available specs can be found in the manual page of
.BR gnutls\-cli (1)
(see the description of the
option
.BR \-\-priority ).

In older versions of GnuTLS, where gnutls\-cli does not support the option
\-\-priority, you can obtain the \(em more limited \(em list of ciphers by calling:

.nf
	gnutls\-cli \-l
.fi

When using Mozilla NSS, the OpenSSL cipher suite specifications are used and
translated into the format used internally by Mozilla NSS.  There isn't an easy
way to list the cipher suites from the command line.  The authoritative list
is in the source code for Mozilla NSS in the file sslinfo.c in the structure
.nf
        static const SSLCipherSuiteInfo suiteInfo[]
.fi
.RE
.TP
.B TLSCACertificateFile <filename>
Specifies the file that contains certificates for all of the Certificate
Authorities that
.B lloadd
will recognize.  The certificate for
the CA that signed the server certificate must be included among
these certificates. If the signing CA was not a top-level (root) CA,
certificates for the entire sequence of CA's from the signing CA to
the top-level CA should be present. Multiple certificates are simply
appended to the file; the order is not significant.
.TP
.B TLSCACertificatePath <path>
Specifies the path of a directory that contains Certificate Authority
certificates in separate individual files. Usually only one of this
or the TLSCACertificateFile is used. This directive is not supported
when using GnuTLS.

When using Mozilla NSS, <path> may contain a Mozilla NSS cert/key
database.  If <path> contains a Mozilla NSS cert/key database and
CA cert files, OpenLDAP will use the cert/key database and will
ignore the CA cert files.
.TP
.B TLSCertificateFile <filename>
Specifies the file that contains the
.B lloadd
server certificate.

When using Mozilla NSS, if using a cert/key database (specified with
TLSCACertificatePath), TLSCertificateFile specifies
the name of the certificate to use:
.nf
	TLSCertificateFile Server-Cert
.fi
If using a token other than the internal built in token, specify the
token name first, followed by a colon:
.nf
	TLSCertificateFile my hardware device:Server-Cert
.fi
Use certutil \-L to list the certificates by name:
.nf
	certutil \-d /path/to/certdbdir \-L
.fi
.TP
.B TLSCertificateKeyFile <filename>
Specifies the file that contains the
.B lloadd
server private key that matches the certificate stored in the
.B TLSCertificateFile
file.  Currently, the private key must not be protected with a password, so
it is of critical importance that it is protected carefully.

When using Mozilla NSS, TLSCertificateKeyFile specifies the name of
a file that contains the password for the key for the certificate specified with
TLSCertificateFile.  The modutil command can be used to turn off password
protection for the cert/key database.  For example, if TLSCACertificatePath
specifies /etc/openldap/certdb as the location of the cert/key database, use
modutil to change the password to the empty string:
.nf
	modutil \-dbdir /etc/openldap/certdb \-changepw 'NSS Certificate DB'
.fi
You must have the old password, if any.  Ignore the WARNING about the running
browser.  Press 'Enter' for the new password.
.TP
.B TLSDHParamFile <filename>
This directive specifies the file that contains parameters for Diffie-Hellman
ephemeral key exchange.  This is required in order to use a DSA certificate on
the server, or an RSA certificate missing the "key encipherment" key usage.
Note that setting this option may also enable
Anonymous Diffie-Hellman key exchanges in certain non-default cipher suites.
Anonymous key exchanges should generally be avoided since they provide no
actual client or server authentication and provide no protection against
man-in-the-middle attacks.
You should append "!ADH" to your cipher suites to ensure that these suites
are not used.
When using Mozilla NSS these parameters are always generated randomly
so this directive is ignored.
.TP
.B TLSECName <name>
Specify the name of a curve to use for Elliptic curve Diffie-Hellman
ephemeral key exchange.  This is required to enable ECDHE algorithms in
OpenSSL.  This option is not used with GnuTLS; the curves may be
chosen in the GnuTLS ciphersuite specification. This option is also
ignored for Mozilla NSS.
.TP
.B TLSProtocolMin <major>[.<minor>]
Specifies minimum SSL/TLS protocol version that will be negotiated.
If the server doesn't support at least that version,
the SSL handshake will fail.
To require TLS 1.x or higher, set this option to 3.(x+1),
e.g.,

.nf
	TLSProtocolMin 3.2
.fi

would require TLS 1.1.
Specifying a minimum that is higher than that supported by the
OpenLDAP implementation will result in it requiring the
highest level that it does support.
This directive is ignored with GnuTLS.
.TP
.B TLSRandFile <filename>
Specifies the file to obtain random bits from when /dev/[u]random
is not available.  Generally set to the name of the EGD/PRNGD socket.
The environment variable RANDFILE can also be used to specify the filename.
This directive is ignored with GnuTLS and Mozilla NSS.
.TP
.B TLSVerifyClient <level>
Specifies what checks to perform on client certificates in an
incoming TLS session, if any.
The
.B <level>
can be specified as one of the following keywords:
.RS
.TP
.B never
This is the default.
.B lloadd
will not ask the client for a certificate.
.TP
.B allow
The client certificate is requested.  If no certificate is provided,
the session proceeds normally.  If a bad certificate is provided,
it will be ignored and the session proceeds normally.
.TP
.B try
The client certificate is requested.  If no certificate is provided,
the session proceeds normally.  If a bad certificate is provided,
the session is immediately terminated.
.TP
.B demand | hard | true
These keywords are all equivalent, for compatibility reasons.
The client certificate is requested.  If no certificate is provided,
or a bad certificate is provided, the session is immediately terminated.
.TP
.B TLSCRLCheck <level>
Specifies if the Certificate Revocation List (CRL) of the CA should be
used to verify if the client certificates have not been revoked. This
requires
.B TLSCACertificatePath
parameter to be set. This directive is ignored with GnuTLS and Mozilla NSS.
.B <level>
can be specified as one of the following keywords:
.RS
.TP
.B none
No CRL checks are performed
.TP
.B peer
Check the CRL of the peer certificate
.TP
.B all
Check the CRL for a whole certificate chain
.RE
.TP
.B TLSCRLFile <filename>
Specifies a file containing a Certificate Revocation List to be used
for verifying that certificates have not been revoked. This directive is
only valid when using GnuTLS and Mozilla NSS.

.SH BACKEND CONFIGURATION
Options in this section describe how the
.B lloadd
connects and authenticates to the backend servers. Backends are organised in groups
.RB ( tiers ).
Backends in the first tier are tried first, if none of them are reachable, the
following tier is tried in the same way. If there is a backend in the tier that
has suitable connections, but they are busy, no further tier is consulted. This
is useful in high availability scenarios where a group of servers (e.g. the
local environment) should be contacted if possible.

It is assumed all backend servers serve the same data. On startup, the
configured connections are set up and those not dedicated to handle bind
requests are authenticated with the backend using the information in the
.B bindconf
option. The authentication configuration is shared between them.
.TP
.B bindconf
.B [bindmethod=simple|sasl]
.B [binddn=<dn>]
.B [saslmech=<mech>]
.B [authcid=<identity>]
.B [authzid=<identity>]
.B [credentials=<passwd>]
.B [realm=<realm>]
.B [secprops=<properties>]
.B [timeout=<seconds>]
.B [network\-timeout=<seconds>]
.B [tcp\-user\-timeout=<milliseconds>]

Specifies the bind credentials
.B lloadd
uses when setting up its regular connections to all backends.

A
.B bindmethod
of
.B simple
requires the options
.B binddn
and
.B credentials
and should only be used when adequate security services
(e.g. TLS or IPSEC) are in place.
.B REMEMBER: simple bind credentials must be in cleartext!
A
.B bindmethod
of
.B sasl
requires the option
.B saslmech.
Depending on the mechanism, an authentication identity and/or
credentials can be specified using
.B authcid
and
.B credentials.
The
.B authzid
parameter may be used to specify an authorization identity.
Specific security properties (as with the
.B sasl\-secprops
keyword above) for a SASL bind can be set with the
.B secprops
option. A non default SASL realm can be set with the
.B realm
option.

The
.B timeout
parameter indicates how long an operation can be pending a response (result,
search entry, ...) from the server in seconds. Due to how timeouts are
detected, the timeout might not be detected and handled up to
.B timeout
seconds after it happens.

The
.B network\-timeout
parameter sets how long the consumer will wait to establish a
network connection to the provider. Once a connection is
established, the
.B timeout
parameter determines how long the consumer will wait for the initial
Bind request to complete.

Timeout set to 0 means no timeout is in effect and by default, no timeouts are
in effect.

The
.B tcp\-user\-timeout
parameter, if non-zero, corresponds to the
.B TCP_USER_TIMEOUT
set on the upstream connections, overriding the operating system setting.
Only some systems support the customization of this parameter, it is
ignored otherwise and system-wide settings are used.

.SH TIER OPTIONS

.TP
.B tier
.B <tier type>

Groups servers which should be considered in the same try. If a viable
connection is found even if busy, the load balancer does not proceed to the
next tier. The process of selection a connection within a tier depends on the
tier's type.

.RE
Available types are:
.TP
.B roundrobin
Servers are tried in order and if one is selected successfully, the following
search will try from the one next on the list.
.TP
.B weighted
Backend servers accept a new option
.B weight=<int>
which indicates how often it should be selected. If unspecified, weight
defaults to 0 and such backends have a slight chance of being selected even
when a non-zero weight backend is configured in the tier. The selection process
is along the lines of
.BR RFC2782 .
.TP
.B bestof
Like with
.BI weighted ,
backends accept the
.B weight=<int>
option. Average latency multiplied by
.B weight
is measured over time. The selection process chooses 2 backends at random,
compares their weighted latencies and the backend with a better (lower) score
is tried. If the backend is not available (or is busy), the other backend is
tried, then backends are chosen in a round-robin order.

Note that unlike
.BI weighted ,
the higher the weight, the higher the "effective" latency and lower the chance
a backend is selected.

.SH BACKEND OPTIONS

.TP
.B backend-server
.B uri=ldap[s]://<hostname>[:port]
.B [retry=<retry interval in ms>]
.B [keepalive=<idle>:<probes>:<interval>]
.B [starttls=yes|critical]
.B [tls_cert=<file>]
.B [tls_key=<file>]
.B [tls_cacert=<file>]
.B [tls_cacertdir=<path>]
.B [tls_reqcert=never|allow|try|demand]
.B [tls_cipher_suite=<ciphers>]
.B [tls_crlcheck=none|peer|all]
.B [tls_protocol_min=<major>[.<minor>]]
.B [numconns=<conns>]
.B [bindconns=<conns>]
.B [max-pending-ops=<ops>]
.B [conn-max-pending=<ops>]

Marks the beginning of a backend definition.

.B uri
specifies the backend as an LDAP URI. If <port> is not given, the standard
LDAP port number (389 or 636) is used.

Lloadd will attempt to maintain
.B numconns
active connections and
.\" unless the
.\" .B vc
.\" feature is enabled,
also
.B bindconns
active connections dedicated to handling client bind requests.

If an error occurs on a working connection, a new connection attempt is
made immediately, if one happens on establishing a new connection to this
backend, lloadd will wait before a new reconnect attempt is made
according to the
.B retry
parameter (default is 5 seconds).

Operations will be distributed across the backend's connections
.RB ( upstreams ).

The parameter
.B conn-max-pending
unless set to
.B 0
(the default), will limit the number unfinished operations per upstream
connection. Similarly,
.B max-pending-ops
will limit the total number or unfinished operations across all backend's
connections,
.BR 0 ,
the default, means no limit will be imposed for this backend.

The
.B keepalive
parameter sets the values of \fIidle\fP, \fIprobes\fP, and \fIinterval\fP
used to check whether a socket is alive;
.I idle
is the number of seconds a connection needs to remain idle before TCP
starts sending keepalive probes;
.I probes
is the maximum number of keepalive probes TCP should send before dropping
the connection;
.I interval
is interval in seconds between individual keepalive probes.
Only some systems support the customization of these values;
the
.B keepalive
parameter is ignored otherwise, and system-wide settings are used.

The
.B starttls
parameter specifies use of the StartTLS extended operation
to establish a TLS session before Binding to the provider. If the
.B critical
argument is supplied, the session will be aborted if the StartTLS request
fails. Otherwise the syncrepl session continues without TLS. The
tls_reqcert setting defaults to "demand" and the other TLS settings
default to the same as the main slapd TLS settings.

.\" .TP
.\" .B readonly on | off
.\" This option puts the backend into "read-only" mode.  Only read
.\" operations (i.e. bind, search, compare) will be directed towards this
.\" backend. By default, readonly is off.
.\" .TP
.\" .B restrict <oplist>
.\" Specify a whitespace separated list of operations that are restricted.
.\" If defined inside a database specification, restrictions apply only
.\" to that database, otherwise they are global.
.\" Operations can be any of
.\" .BR add ,
.\" .BR bind ,
.\" .BR compare ,
.\" .BR delete ,
.\" .BR extended[=<OID>] ,
.\" .BR modify ,
.\" .BR rename ,
.\" .BR search ,
.\" or the special pseudo-operations
.\" .B read
.\" and
.\" .BR write ,
.\" which respectively summarize read and write operations.
.\" The use of
.\" .I restrict write
.\" is equivalent to
.\" .I readonly on
.\" (see above).
.\" The
.\" .B extended
.\" keyword allows one to indicate the OID of the specific operation
.\" to be restricted.

.SH EXAMPLES
.LP
Here is a short example of a configuration file:
.LP
.RS
.nf
argsfile  LOCALSTATEDIR/run/lloadd.args
pidfile   LOCALSTATEDIR/run/lloadd.pid

# cancel not supported yet
restrict_exop 1.3.6.1.1.8 reject

# turn not supported
restrict_exop 1.3.6.1.1.19 reject

# TXN Exop if desired, otherwise reject
restrict_exop 1.3.6.1.1.21.1 connection

# Paged results control
restrict_control 1.2.840.113556.1.4.319 connection

# VLV control
restrict_control 2.16.840.1.113730.3.4.9 connection

bindconf
    bindmethod=simple
    binddn=cn=test
    credentials=pass

tier weighted
backend-server
    uri=ldap://ldap1.example.com
    numconns=3
    bindconns=2
    retry=5000
    max-pending-ops=5
    conn-max-pending=3
    weight=5

backend-server
    uri=ldap://ldap2.example.com
    numconns=3
    bindconns=2
    retry=5000
    max-pending-ops=5
    conn-max-pending=3
    weight=10
.fi
.RE
.LP
"OpenLDAP Administrator's Guide" contains a longer annotated
example of a configuration file.
The original ETCDIR/lloadd.conf is another example.

.SH LIMITATIONS
Support for proxying SASL Binds is limited to the
.B EXTERNAL
mechanism (and only to extract the DN of a client TLS cerificate if used during
the last renegotiation) and mechanisms that rely neither on connection metadata
(as Kerberos does) nor establish a SASL integrity/confidentialiy layer (again,
some Kerberos mechanisms,
.B DIGEST-MD5
can negotiate this).

.SH FILES
.TP
ETCDIR/lloadd.conf
default lloadd configuration file
.SH SEE ALSO
.BR ldap (3),
.BR gnutls\-cli (1),
.BR slapd.conf (5),
.BR tcp (7),
.BR lloadd (8),
.BR slapd (8).
.LP
"OpenLDAP Administrator's Guide" (http://www.OpenLDAP.org/doc/admin/)
.SH ACKNOWLEDGEMENTS
.so ../Project