Scroll to navigation

YAWS.CONF(5) User Commands YAWS.CONF(5)

NAME

/etc/yaws/yaws.conf - Configuration file for the Yaws web server

DESCRIPTION

Yaws is fast lightweight web server. It reads a configuration file called yaws.conf to control its operations. The configuration contains two distinct parts: a global part which affects all the virtual hosts and a server part where options for each virtual host is supplied.

GLOBAL PART

All Yaws logs will be written to files in this directory. If specified with +, Yaws will attempt to create the directory if it does not exist. There are several different log files written by Yaws:

report.log - this is a text file that contains all error logger printouts from Yaws.

<Host>.access - for each virtual host served by Yaws, a file <Host>.access will be written that contains an access log in NCSA combined/XLF/ELF log format. (See http://www.w3.org/TR/WD-logfile.html for more details on Extended Log File Format.)

<Host>.auth - for each virtual host served by Yaws, a file <Host>.auth will be written which contains all http auth related messages.

trace_<YYYYMMDD_hhmmss> - Trace files are written in this subdirectory, suffixed by the creation date.

trace.<Pid>.http - this file contains the HTTP trace if that is enabled, where <Pid> is the process id handling the TCP connection.

trace.<Pid>.traffic - this file contains the traffic trace if that is enabled, where <Pid> is the process id handling the TCP connection.

Note that <Host>.access and <Host>.auth files will be used only if the directive logger_mod is not set or set to yaws_log. The default value for logdir is "."

This directive adds Directory to the Erlang search path. It is possible to have several of these commands in the configuration file. The default value is "yaws_dir"/examples/ebin

This directive defines a Directory as a source directory. Yaws will compile all erlang modules found in this directory and all its subdirectories. The compilation occurs when the configuration is loaded or reloaded. The include_dir directives are used to search for includes files. Multiple src_dir directives may be used. There is no such directory configured by default.

It is possible to run multiple Yaws servers on the same machine. We use the id of a Yaws server to control it using the different control commands such as:


# /usr/local/bin/yaws --id foobar --stop

To stop the Yaws server with id "foobar". Each Yaws server will write its internal data into a file called $HOME/.yaws/yaws/ID where ID is the identity of the server. Yaws also creates a file called $HOME/.yaws/yaws/ID/CTL which contains the port number where the server is listening for control commands. The default id is "default".

This directive sets the "Server: " output header to the custom value. The default value is "yaws/%VSN%, Yet Another Web Server".

This directive adds Directory to the path of directories where the Erlang compiler searches for include files. We need to use this if we want to include .hrl files in our Yaws Erlang code. It is possible to have several of these commands in the configuration file. The default value is "yaws_dir"/examples/include.

Yaws will cache small files such as commonly accessed GIF images in RAM. This directive sets a maximum number on the number of cached files. The default value is 400.

This directive controls the total amount of RAM which can maximally be used for cached RAM files. The default value is 1000000, 1 megabyte.

This directive sets a maximum size on the files that are RAM cached by Yaws. The default value is 8000, 8 kBytes.

The RAM cache is used to serve pages that sit in the cache. An entry sits in cache at most cache_refresh_secs number of seconds. The default is 30. This means that when the content is updated under the docroot, that change doesn't show until 30 seconds have passed. While developing a Yaws site, it may be convenient to set this value to 0. If the debug flag (-d) is passed to the Yaws start script, this value is automatically set to 0.

This enables traffic or http tracing. Tracing is also possible to enable with a command line flag to Yaws. Default is false.

Deprecated and ignored. Now, this target must be set in server part.

Set this value to control the maximum number of connections from HTTP clients into the server. This is implemented by closing the last socket if the limit threshold is reached.

Normally, Yaws does not restrict the number of times a connection is kept alive using keepalive. Setting this parameter to an integer X will ensure that connections are closed once they have been used X times. This can be a useful to guard against long running connections collecting too much garbage in the Erlang VM.

Set process spawn options for client acceptor processes. Options must be specified as a quoted string of either the atom undefined or as a proplist of valid process options. The supported options are fullsweep_after, min_heap_size, and min_bin_vheap_size, each taking an associated integer value. Other process options are ignored. The proplist may also be empty. See erlang:spawn_opt/4 for details on these options.

Set the chunk size used by Yaws to send large files. The default value is 10240.

Set the version of sendfile method to use to send large files:

erlang - use file:sendfile/5.

disable - use gen_tcp:send/2.

The default value is erlang.

Set the size of the pool of cached acceptor processes. The specified value must be greater than or equal to 0. The default value is 8. Specifying a value of 0 effectively disables the process pool.

The logs written by Yaws are all wrap logs, the default value at the size where they wrap around and the original gets renamed to File.old is 1000000, 1 megabyte. This value can be changed.
If we set the value to 0 the logs will never wrap. If we want to use Yaws in combination with a more traditional log wrapper such as logrotate, set the size to 0 and Yaws will reopen the logfiles once they have be renamed/removed.

By default the client host IP is not resolved in the access logs.

Fail completely or not if Yaws fails to bind a listen socket Default is true.

If true, a soap server will be started at startup of Yaws. Default is false.

If enable_soap is true, a startup Yaws will invoke yaws_soap_srv:setup() to setup modules set here. ModuleSetting is either a triad like <Mod, HandlerFunc, WsdlFile> or a quadruple form like <Mod, HandlerFunc, WsdlFile, Prefix> which specifies the prefix. A prefix will be used as argument of yaws_soap_lib:initModel() and then be used as a XML namespace prefix. Note, the WsdlFile here should be an absolute-path file in local file systems.

For example, we can specify


soap_srv_mods=<Mod1, Handler, Wsdl1> <Mod2, Handler, Wsdl2, Prefix> ...

this target is deprecated and useless. use 'php_handler' target in server part instead.
The name of (and possibly path to) the php executable used to interpret php scripts (if allowed). Default is php_exe_path = php-cgi.

Enable or disable copying of the error log. When we run in embedded mode, there may very well be some other systems process that is responsible for writing the errorlog to a file whereas when we run in normal standalone mode, we typically want the Erlang errorlog written to a report.log file. Default value is true.

Allows specifying a different Yaws session storage mechanism instead of an ETS table. One of the drawbacks of the default yaws_session_server implementation is that server side cookies are lost when the server restarts. Specifying a different module here will pass all write/read operations to this module (it must implement appropriate callbacks).

Allows specifying a different Yaws session cookie generator than the built-in default. Module is expected to provide a new_cookie/0 function that returns a session cookie in the form of a list. Such a cookie generator module must be careful to return a unique cookie each time it's called.

Controls Yaws session idle cleanup. If a server has been idle for ysession_idle_timeout milliseconds, check all Yaws sessions and remove any that have timed out. The default ysession_idle_timeout value is 2*60*1000 (2 minutes).

Controls Yaws session periodic cleanup. Every ysession_long_timeout milliseconds, check all Yaws sessions and remove any that have timed out. The default ysession_long_timeout value is 60*60*1000 (1 hour).

At startup Yaws will invoke ModuleName:start() in a separate process. It is possible to have several runmods. This is useful if we want to reuse the Yaws startup shell script for our own application.

When Yaws gets a request, it extracts the Host header from the client request to choose a virtual server amongst all servers with the same IP/Port pair. This configuration parameter decides whether Yaws should pick the first server (as defined in the yaws.conf file) if no name match or not. If this is false and no Host header is present in the request, Yaws returns a 400 Bad Request as required by the HTTP standard. In real live hosting scenarios we typically want this to be false, whereas in testing/development scenarios it may be convenient to set it to true. Default is true.

If the HTTP session will be kept alive (i.e., not immediately closed) it will close after keepalive_timeout milliseconds unless a new request is received in that time. The default value is 30000. The value infinity is legal but not recommended.

Load specified config file. Absolute paths or relative ones to the configuration location are allowed. Unix-style wildcard strings can be used to include several files at once. See filelib:wildcard/1 for details. Hidden files, starting by a dot, will be ignored. For example:


subconfig = /etc/yaws/global.conf
subconfig = /etc/yaws/vhosts/*.conf

Or, relatively to the configuration location:


subconfig = global.conf
subconfig = vhosts/*.conf

Load all config files found in the specified directory. The given Directory can be an absolute path or relative to the configuration location. Hidden files, starting by a dot, will be ignored.

this target is deprecated and will be ignored.

Defines the default MIME type to be used where Yaws cannot determine it by its MIME types mappings. Default is text/plain.

Defines the default charset to be added when a response content-type is text/*. By default, no charset is added.

Overrides the default mime.types file included with Yaws. This file must use the following format:


# Lines beginning with a '#' or a whitespace are ignored
# blank lines are also ignored
<MIME type> <space separated file extensions>

The default file is located at ${PREFIX}/lib/yaws/priv/mime.types. You should not edit this file because it may be replaced when you upgrade your server.

Specifies one or more mappings between MIME types and file extensions. More than one extension can be assigned to a MIME type. ListOfTypes is defined as follows:


add_types = <MimeType1, Ext> <MimeType2, Ext1 Ext2 ...> ...

The mappings defined using this directive will overload all other definitions. If a file extension is defined several times, only the last one is kept. Multiple add_types directives may be used.

Specifies one or more mappings between charsets and file extensions. More than one extension can be assigned to a charset. ListOfCharsets is defined as follows:


add_charsets = <Charset1, Ext> <Charset2, Ext1 Ext2 ...> ...

The mappings defined using this directive will overload all other definitions. If a file extension is defined several times, only the last one is kept. Multiple add_charsets directives may be used.

Enables or disables the TLS SNI (Server Name Indication) support.

When disabled (or not supported), all virtual servers in the same group (same IP/Port) must share the same SSL configuration, especially the same SSL certificate. Only the HTTP Host header will be considered to find the right virtual server.

When enabled, SSL configuration can be different from one virtual server to another; each one can have its own SSL certificate. In this case, if a client provides a SNI hostname, it will be used to find the right virtual server. To accept the SNI information from the client, the first virtual server -- the default one, see pick_first_virthost_on_nomatch -- must include TLS as a permitted protocol.

If the sni directive is set to enable, non-SNI clients are allowed. For such clients, virtual servers are selected as if Yaws did not have SNI support. If it is set to strict, SNI hostname is mandatory to access a SSL virtual server. But in all cases, when SNI support is enabled, if a client provides a SNI hostname, it must match the HTTP Host header (which is mandatory too). Note that the first virtual server (the default one) will be used for any request where the provided SNI hostname doesn't match any of virtual server names. So, it is important that the first virtual server have the most restrictive access control, otherwise clients can access restricted resources by sending a request for any unknown hostname. (This isn't actually any different from using virtual servers without SNI support.) If you're using self-signed certificates, be sure to also set the depth configuration variable to 0 to avoid following certificate chains.

The sni directive is a global one, so if you set it to strict, non-SNI clients will be refused for all SSL groups. See require_sni directive from the server part to mitigate this requirement.

Default is disable.

SERVER PART

Yaws can virthost several web servers on the same IP address as well as several web servers on different IP addresses. This includes SSL servers.

Each virtual host is defined within a matching pair of <server ServerName> and </server>. The ServerName will be the name of the webserver.

The following directives are allowed inside a server definition.

This makes the server listen on Port. Default is 8000.
This makes the server listen on IpAddress. When virthosting several servers on the same ip/port address, if the browser doesn't send a Host: field, Yaws will pick the first server specified in the config file. If the specified IP address is 0.0.0.0 Yaws will listen on all local IP addresses on the specified port. Default is 127.0.0.1. Multiple listen directives may be used to specify several addresses to listen on.

This sets the TCP listen backlog for the server to define the maximum length the queue of pending connections may grow to. The default is 1024.

<listen_opts> ... </listen_opts>
Defines extra options to be set on the listen socket and, by inheritance, on accepted sockets. See inet:setopts/2 for details. Supported options are:

buffer = Integer (default: same as inet:setopts/2)

delay_send = true | false (default: same as inet:setopts/2)

linger = Integer | false (default: same as inet:setopts/2)

nodelay = true | false (default: same as inet:setopts/2)

priority = Integer (default: same as inet:setopts/2)

sndbuf = Integer (default: same as inet:setopts/2)

recbuf = Integer (default: same as inet:setopts/2)

send_timeout = Integer | infinity (default: same as inet:setopts/2)

send_timeout_close = true | false (default: same as inet:setopts/2)

This directive sets the "Server: " output header to the custom value and overloads the global one for this virtual server.

Same as subconfig directive of the global part, but here files should only contain directives allowed in the server part.

Same as subconfigdir directive of the global part, but here files should only contain directives allowed in server part.

This forces all local redirects issued by the server to go to Host. This is useful when Yaws listens to a port which is different from the port that the user connects to. For example, running Yaws as a non-privileged user makes it impossible to listen to port 80, since that port can only be opened by a privileged user. Instead Yaws listens to a high port number port, 8000, and iptables are used to redirect traffic to port 80 to port 8000 (most NAT:ing firewalls will also do this for you).

This forces all local redirects issued by the server to use this method. This is useful when an SSL off-loader, or stunnel, is used in front of Yaws.

Enable or disable the auth log for this virtual server. Default is true.

Setting this directive to false turns off traffic logging for this virtual server. The default value is true.

It is possible to set a special module that handles access and auth logging. The default is to log all web server traffic to <Host>.access and <Host>.auth files in the configured or default logdir.
This module must implement the behaviour yaws_logger. Default value is yaws_log.

The following functions should be exported:

Module:open_log(ServerName, Type, LogDir)

When Yaws is started, this function is called for this virtual server. If the initialization is successful, the function must return {true,State} and if an error occurred, it must return false.

Module:close_log(ServerName, Type, State)
This function is called for this virtual server when Yaws is stopped.

Module:wrap_log(ServerName, Type, State, LogWrapSize)
This function is used to rotate log files. It is regularly called by Yaws and must return the possibly updated internal NewState.

Module:write_log(ServerName, Type, State, Infos)
When it needs to log a message, Yaws will call this function. The parameter Infos is {Ip,Req,InHdrs,OutHdrs,Time} for an access log and {Ip,Path,Item} for an auth log, where:

Ip - IP address of the accessing client (as a tuple).

Req - the HTTP method, URI path, and HTTP version of the request (as a #http_request{} record).

InHdrs - the HTTP headers which were received from the WWW client (as a #headers{} record).

OutHdrs - the HTTP headers sent to the WWW client (as a #outh{} record)

Path - the URI path of the request (as a string).

Item - the result of an authentication request. May be {ok,User}, 403 or {401,Realm}.

Time - The time taken to serve the request, in microseconds.

For all of these callbacks, ServerName is the virtual server's name, Type is the atom access or auth and State is the internal state of the logger.

Defines a module to control access to this virtual server. Access can be controlled based on the IP address of the client. It is also possible to throttle HTTP requests based on the client's download rate. This module must implement the behaviour yaws_shaper.

There is no such module configured by default.

Setting this directive to false disallows the automatic dir listing feature of Yaws. A status code 403 Forbidden will be sent. Set to true_nozip to avoid the auto-generated all.zip entries. Default is false.

Add additional CGI or FastCGI variables. For example:


<extra_cgi_vars dir='/path/to/some/scripts'>
var = val
...
</extra_cgi_vars>

Turns on/off statistics gathering for a virtual server. Default is false.

The hostname and TCP port number of a FastCGI application server. To specify an IPv6 address, put it inside square brackets (ex: "[::1]:9000"). The TCP port number is not optional. There is no default value.

Enable or disable tracing of FastCGI protocol messages as info log messages. Disabled by default.

Enable or disable logging of application error messages (output to stderr and non-zero exit value). Disabled by default.

Turns on or off deflate compression for a server. Default is false.

<deflate> ... </deflate>
This begins and ends the deflate compression configuration for this server. The following items are allowed within a matching pair of <deflate> and </deflate> delimiters.

min_compress_size = nolimit | Integer

Defines the smallest response size that will be compressed. If nolimit is not used, the specified value must be strictly positive. The default value is nolimit.

compression_level = none | default | best_compression | best_speed | 0..9
Defines the compression level to be used. 0 (none), gives no compression at all, 1 (best_speed) gives best speed and 9 (best_compression) gives best compression. The default value is default.

window_size = 9..15
Specifies the zlib compression window size. It should be in the range 9 through 15. Larger values of this parameter result in better compression at the expense of memory usage. The default value is 15.

mem_level = 1..9
Specifies how much memory should be allocated for the internal compression state. mem_level=1 uses minimum memory but is slow and reduces compression ratio; mem_level=9 uses maximum memory for optimal speed. The default value is 8.

strategy = default | filtered | huffman_only
This parameter is used to tune the compression algorithm. See zlib(3erl) for more details on the strategy parameter. The default value is default.

use_gzip_static = true | false
If true, Yaws will try to serve precompressed versions of static files. It will look for precompressed files in the same location as original files that end in ".gz". Only files that do not fit in the cache are concerned. The default value is false.

mime_types = ListOfTypes | defaults | all
Restricts the deflate compression to particular MIME types. The special value all enable it for all types (It is a synonym of `*/*'). MIME types into ListOfTypes must have the form `type/subtype' or `type/*' (indicating all subtypes of that type). Here is an example:


mime_types = default image/*
mime_types = application/xml application/xhtml+xml application/rss+xml

By default, the following MIME types are compressed (if deflate is set to true): text/*, application/rtf, application/msword, application/pdf, application/x-dvi, application/javascript. Multiple mime_types directives can be used.

This makes the server serve all its content from Directory.
It is possible to pass a space-separated list of directories as docroot. If this is the case, the various directories will be searched in order for the requested file. This also works with the ssi and yssi constructs where the full list of directories will be searched for files to ssi/yssi include. Multiple docroot directives can be used. You need at least one valid docroot, invalid docroots are skipped with their associated auth structures.

If true, the docroot will not be searched for .yaws_auth files. This is useful when the docroot is quite large and the time to search it is prohibitive when Yaws starts up. Defaults to false.

When a Yaws file receives large POSTs, the amount of data received in each chunk is determined by this parameter. The default value is 10240. Setting it to nolimit is potentially dangerous.

Turns on the DAV protocol for this server. The dav support in Yaws is highly limited. If dav is turned on, .yaws processing of .yaws pages is turned off. Default is false. The socket read timeout is supplied by the keepalive_timeout setting. If the read is not done within the timeout, the POST will fail.

If this value is set to false Yaws will never do tilde expansion. The default is false. tilde_expansion is the mechanism whereby a URL on the form http://www.foo.com/~username is changed into a request where the docroot for that particular request is set to the directory ~username/public_html/.

The allowed script types for this server. Recognized are `yaws', `cgi', `fcgi', `php'. Default is allowed_scripts = yaws php cgi fcgi.

Note: for fcgi scripts, the FastCGI application server is only called if a local file with the .fcgi extension exists. However, the contents of the local .fcgi file are ignored.

The allowed script types for this server when executing files in a users public_html folder Recognized are `yaws', `cgi', `fcgi', `php'. Default is tilde_allowed_scripts = i.e. empty

This directive sets the list of resources to look for, when a directory is requested by the client. If the last entry begins with a `/', and none of the earlier resources are found, Yaws will perform a redirect to this uri. Default is index_files = index.yaws index.html index.php.

If any of the names in ListOfModuleNames appear as components in the path for a request, the path request parsing will terminate and that module will be called. There is also an alternate syntax for specifying the appmods if we don't want our internal erlang module names to be exposed in the URL paths. We can specify


appmods = <Path1, Module1> <Path2, Modules2> ...

Assume for example that we have the URL http://www.hyber.org/myapp/foo/bar/baz?user=joe while we have the module foo defined as an appmod, the function foo:out(Arg) will be invoked instead of searching the filesystems below the point foo.
The Arg argument will have the missing path part supplied in its appmoddata field.

It is also possible to exclude certain directories from appmod processing. This is particulaly interesting for '/' appmods. Here is an example:


appmods = </, myapp exclude_paths icons js top/static>

The above configuration will invoke the 'myapp' erlang module on everything except any file found in directories 'icons', 'js' and 'top/static' relative to the docroot.

Set DispatchModule as a server-specific request dispatching module. Yaws expects DispatchModule to export a dispatch/1 function. When it receives a request, Yaws passes an #arg{} record to the dispatch module's dispatch/1 function, which returns one of the following atom results:

done - this indicates the dispatch module handled the request itself and already sent the response, and Yaws should resume watching for new requests on the connection

closed - same as done but the DispatchModule also closed the connection

continue - the dispatch module has decided not to handle the request, and instead wants Yaws to perform its regular request dispatching

Note that when DispatchModule handles a request itself, Yaws does not support tracing, increment statistics counters or allow traffic shaping for that request. It does however still keep track of maximum keepalive uses on the connection.

It is possible to set a special module that handles 404 Not Found messages. The function Module:out404(Arg, GC, SC) will be invoked. The arguments are

Arg - a #arg{} record

GC - a #gconf{} record (defined in yaws.hrl)

SC - a #sconf{} record (defined in yaws.hrl)

The function can and must do the same things that a normal out/1 does.

It is possible to set a special module that handles 401 Unauthorized messages. This can for example be used to display a login page instead. The function Module:out401(Arg, Auth, Realm) will be invoked. The arguments are

Arg - a #arg{} record

Auth - a #auth{} record

Realm - a string

The function can and must do the same things that a normal out/1 does.

It is possible to set a special module that handles the HTML generation of server crash messages. The default is to display the entire formatted crash message in the browser. This is good for debugging but not in production.
The function Module:crashmsg(Arg, SC, Str) will be called. The Str is the real crash message formatted as a string.
The function must return, {content,MimeType,Cont} or {html, Str} or {ehtml, Term}. That data will be shipped to the client.

Controls the setting of the Expires HTTP header and the max-age directive of the Cache-Control HTTP header in server responses for specific MIME types. The expiration date can be set as relative to either the time the source file was last modified; as the time of the client access; or as always in order to empty the cache altogether. ListOfExpires is defined as follows:


expires = <MimeType1, access+Seconds> <MimeType2, modify+Seconds> <MimeType3, always> ...

A MimeType can also have a wildcard as subtype or both as subtype and type, like type/* or */*.

These HTTP headers are an instruction to the client about the document's validity and persistence. If cached, the document may be fetched from the cache rather than from the source until this time has passed. After that, the cache copy is considered "expired" and invalid, and a new copy must be obtained from the source. Here is an example:


expires = <image/gif, access+2592000> <image/png, access+2592000>
expires = <image/jpeg, access+2592000> <text/css, access+2592000>
expires = <text/*, always>

and here is another:


expires = <*/*, always>

It is possible to install a module that rewrites all the Arg #arg{} records at an early stage in the Yaws server. This can be used to do various things such as checking a cookie, rewriting paths etc. An arg_rewrite_mod must export an arg_rewrite/1 function taking and returning an #arg{} record. If the function wants to return a response, it must set the #arg.state field of its return value to an instance of the #rewrite_response{} record.

The module yaws_vdir can be used in case you want to serve static content that is not located in your docroot. See the example at the bottom of this man page for how to use the opaque + vdir elements to instruct the yaws_vdir module what paths to rewrite.

Defines a user provided callback module. At startup of the server, Module:start/1 will be called. The #sconf{} record (defined in yaws.hrl) will be used as the input argument. This makes it possible for a user application to synchronize the startup with the Yaws server as well as getting hold of user specific configuration data, see the explanation for the <opaque> context.

Make Yaws a reverse proxy. Prefix is a path inside our own docroot and Url argument is a URL pointing to a website we want to "mount" under the Prefix path. This example:


revproxy = /tmp/foo http://yaws.hyber.org

makes the hyber website appear under /tmp/foo.

It is possible to have multiple reverse proxies inside the same server.

You can optionally configure an interception module for each reverse proxy, allowing your application to examine and modify requests and HTTP headers as they pass through the proxy from client to backend server and also examine and modify responses and HTTP headers as they return from the backend server through the proxy to the client.

You specify an interception module by including the optional intercept_mod keyword followed by Module, which should be the name of your interception module.

An interception module is expected to export two functions: rewrite_request/2 and rewrite_response/2. The two arguments passed to rewrite_request/2 function are a #http_request{} record and a #headers{} record, whereas rewrite_response/2 function takes a #http_response{} record and also a #headers{} record. You can find definitions for these record types in the yaws_api.hrl header file. Each function can examine each record instance and can either return each original instance or can return a modified copy of each instance in its response. The rewrite_request/2 function should return a tuple of the following form:


{ok, #http_request{}, #headers{}}

and the rewrite_response/2 function should similarly return a tuple of the following form:


{ok, #http_response{}, #headers{}}

A #headers{} record can easily be manipulated in an interceptor using the functions listed below:


yaws_api:set_header/2, yaws_api:set_header/3
yaws_api:get_header/2, yaws_api:get_header/3
yaws_api:delete_header/2

Any failures in your interception module's functions will result in HTTP status code 500, indicating an internal server error.

Make Yaws a forward proxy. By enabling this option you can use Yaws as a proxy for outgoing web traffic, typically by configuring the proxy settings in a web-browser to explicitly target Yaws as its proxy server.

If we're virthosting several servers and want to force a server to match specific Host: headers we can do this with the "servername" directive. This name doesn't necessarily have to be the same as the the name inside <server Name> in certain NAT scenarios. Rarely used feature.

This directive sets the alternate names for a virtual host. A server alias may contain wildcards:


'*' matches any sequence of zero or more characters
'?' matches one character unless that character is a period ('.')
Multiple serveralias directives may be used. Here is an example:


<server server.domain.com>
serveralias = server server2.domain.com server2
serveralias = *.server.domain.com *.server?.domain.com
...
</server>

Set handler to interpret .php files. It can be one of the following definitions:

php_handler = <cgi, Filename> - The name of (and possibly path to) the php executable used to interpret php scripts (if allowed).

php_handler = <fcgi, Host:Port> - Use the specified fastcgi server to interpret .php files (if allowed).

Yaws does not start the PHP interpreter in fastcgi mode for you. To run PHP in fastcgi mode, call it with the -b option. For example:


php5-cgi -b '127.0.0.1:54321'

This starts a php5 in fastcgi mode listening on the local network interface. To make use of this PHP server from Yaws, specify:


php_handler = <fcgi, 127.0.0.1:54321>

If you need to specify an IPv6 address, use square brackets:


php_handler = <fcgi, [::1]:54321>

The PHP interpreter needs read access to the files it is to serve. Thus, if you run it in a different security context than Yaws itself, make sure it has access to the .php files.
Please note that anyone who is able to connect to the php fastcgi server directly can use it to read any file to which it has read access. You should consider this when setting up a system with several mutually untrusted instances of php.

php_handler = <extern, Module:Function | Node:Module:Function> - Use an external handler, possibly on another node, to interpret .php files (if allowed).

To interpret a .php file, the function Module:Function(Arg) will be invoked (Evaluated inside a rpc call if a Node is specified), where Arg is an #arg{} record.
The function must do the same things that a normal out/1 does.

Default value is <cgi, "/usr/bin/php-cgi">.

this target is deprecated. use 'php_handler' target in server part instead.
Using this directive is the same as: php_handler = <fcgi, Host:Port>.

Overloads the global default_type value for this virtual server.
Overloads the global default_charset value for this virtual server.

Overloads the global mime_type_file value for this virtual server. Mappings defined in File will not overload those defined by add_types directives in the global part.

Overloads the global add_types values for this virtual server. If a mapping is defined in the global part and redefined in a server part using this directive, then it is replaced. Else it is kept.

Overloads the global add_charsets values for this virtual server. If a mapping is defined in the global part and redefined in a server part using this directive, then it is replaced. Else it is kept.

For fcgi servers and revproxy URLs, define the name resolution preference. For example, to perform only IPv4 name resolution, use [inet]. To do both IPv4 and IPv6 but try IPv6 first, use [inet6, inet]. Default value is [inet].

<ssl> ... </ssl>

This begins and ends an SSL configuration for this server. It's possible to virthost several SSL servers on the same IP/Port. If SNI support is disabled or not supported, they must share the same certificate configuration. In this situation, it is complicated to virthost several SSL servers on the same IP/Port since the certificate is typically bound to a domainname in the common name part of the certificate. One solution to this problem is to have a certificate with multiple subjectAltNames. If SNI support is enabled, SSL servers on the same IP/Port can have their own SSL configuration with a different SSL certificate for each one. See the global sni directive.

The SNI support was introduced in the SSL application in Erlang/OTP 18.0. It is an extension to the TLS protocol (RFC 4366), which allows the client to include the requested hostname in the first message of its SSL handshake.

See also http://wiki.cacert.org/VhostTaskForce#Interoperability_Test for browser compatibility.

keyfile = File

Specifies which file contains the private key for the certificate. If not specified then the certificate file will be used.

certfile = File
Specifies which file contains the certificate for the server.

cacertfile = File
A file containing trusted certificates to use during client authentication and to use when attempting to build the server certificate chain. The list is also used in the list of acceptable client CAs passed to the client when a certificate is requested.

dhfile = File
A file containing PEM-encoded Diffie-Hellman parameters to be used by the server if a cipher suite using Diffie-Hellman key exchange is negotiated. If not specified, default parameters are used.

verify = verify_none | verify_peer
Specifies the level of verification the server does on client certs. Setting verify_none means that the x509 validation will be skipped (no certificate request is sent to the client), verify_peer means that a certificate request is sent to the client (x509 validation is performed.

You might want to use fail_if_no_peer_cert in combination with verify_peer.

fail_if_no_peer_cert = true | false
If verify is set to verify_peer and set to true the connection will fail if the client does not send a certificate (i.e. an empty certificate). If set to false the server will fail only if an invalid certificate is supplied (an empty certificate is considered valid).

depth = Int
Specifies the depth of certificate chains the server is prepared to follow when verifying client certs. For the OTP new SSL implementation it is also used to specify how far the server (Yaws in our case) shall follow the SSL certificates we present to the clients. Hence, using self-signed certs, we typically need to set this to 0.

password = String
If the private key is encrypted on disc, this password is the 3DES key to decrypt it.

ciphers = String
This string specifies the SSL cipher string. The syntax of the SSL cipher string is a 4-tuple representation of the map returned by ssl:cipher_suites/2,3: {#{key_exchange}, #{cipher}, #{mac}, #{prf}}.
ciphers = "[{dhe_rsa,aes_256_cbc,sha,default_prf}, \

{dhe_dss,aes_256_cbc,sha,default_prf}]"

In older versions of Yaws, a cipher tuple lacked the #{prf} element. When Yaws reads a cipher of the old format from configuration, it attempts to convert it to a 4-tuple by adding default_prf for the #{prf} element. Be aware that this may not work for all ciphers; if it fails, manual intervention is needed to properly configure the ciphers in the new format.

eccs = String
This string specifies the supported Elliptic Curve Cryptography (ECC). It must be a subset of ssl:eccs(). For PCI DSS compliance (which is the main reason why you would want to change this), set it on a single line to:
eccs = "[sect571r1, sect571k1, secp521r1, brainpoolP512r1, \

sect409k1, sect409r1, brainpoolP384r1, secp384r1, \
sect283k1, sect283r1, brainpoolP256r1, secp256k1, \
secp256r1, sect239k1, xsect233k1, sect233r1, \
secp224k1, secp224r1]"

secure_renegotiate = true | false | undefined
Specifies whether to reject renegotiation attempt that does not live up to RFC 5746. By default secure_renegotiate is set to false for protocol versions that support it, i.e. secure renegotiation will be used if possible but it will fallback to unsecure renegotiation if the peer does not support RFC 5746. Set it to undefined to use the ssl module default setting to avoid errors with protocol versions that don't support it, such as TLS version 1.3. For more details, see the ssl manual page at http://www.erlang.org/doc/man/ssl.html

client_renegotiation = true | false | undefined
Enables or disables the Erlang/OTP SSL application client renegotiation option. Defaults to true for protocol versions that support it. Set it to undefined to use the ssl module default setting to avoid errors with protocol versions that don't support it, such as TLS version 1.3. For more details, see the ssl manual page at http://www.erlang.org/doc/man/ssl.html

honor_cipher_order = true | false
If true (the default), use the server's preference for cipher selection. If false, use the client's preference.

protocol_version = ProtocolList
Specifies the list of SSL protocols that will be supported. If not set, defaults to all protocols supported by the erlang ssl application. For example, to support only TLS versions 1.3, 1.2, 1.1, and 1:
protocol_version = tlsv1.3, tlsv1.2, tlsv1.1, tlsv1

require_sni = true | false
If true,the server will reject non-SNI clients and clients providing an unknown SNI hostname (this last remark is only relevant for the first virtual server of a SSL group). This directive is ignored if SNI support is disabled (or not supported).

Default is false.

<redirect> ... </redirect>
Defines a redirect mapping. The following items are allowed within a matching pair of <redirect> and </redirect> delimiters.

We can have a series of redirect rules in one of the formats below:


Path = URL
Path = code
Path = code URL

Path must be an url-decoded path beginning with a slash. URL may be either a relative URL (a path beginning with a slash), or an absolute URL. In the first case, the scheme:hostname:port of the current server will be added. All accesses to Path will be redirected to URL/Path (or scheme:hostname:port/URL/Path if URL is relative). URL must be url-encoded. Note that the original path is appended to the redirected URL.

For example, assume we have the following redirect configuration:


<redirect>
/foo = http://www.mysite.org/zapp
/bar = /tomato.html
</redirect>

Assuming this config resides on a site called http://abc.com, we have the following redirects:


http://abc.com/foo -> http://www.mysite.org/zapp/foo
http://abc.com/foo/test -> http://www.mysite.org/zapp/foo/test
http://abc.com/bar -> http://abc.com/tomato.html/bar
http://abc.com/bar/x/y/z -> http://abc.com/tomato.html/bar/x/y/z

By default, Yaws will perform a 302 redirect. The HTTP status code can be changed using the code parameter. Note that the status code must be known by Yaws.

For 3xx status codes, the URL parameter must be present and will be used to build the new location.
For other status codes (1xx, 2xx, 4xx and 5xx), it can be omitted. In the absence of URL, Yaws will return a generic response with the specified status code.
Otherwise, the URL parameter must be a relative URL and will be used to customize the response.

Sometimes we do not want to have the original path appended to the redirected path. To get that behaviour we specify the config with '==' instead of '='.


<redirect>
/foo == http://www.mysite.org/zapp
/bar = /tomato.html
</redirect>

Now a request for http://abc.com/foo/x/y/z simply gets redirected to http://www.mysite.org/zapp. This is typically used when we simply want a static redirect at some place in the docroot.

When we specify a relative URL as the target for the redirect, the redirect will be to the current http(s) server.

<auth> ... </auth>
Defines an auth structure. The following items are allowed within a matching pair of <auth> and </auth> delimiters.

docroot = Docroot

If a docroot is defined, this auth structure will be tested only for requests in the specified docroot. No docroot configured means all docroots. If two auth structures are defined, one with a docroot and one with no docroot, the first of both overrides the second one for requests in the configured docroot.

dir = Dir
Makes Dir to be controlled by WWW-authenticate headers. In order for a user to have access to WWW-Authenticate controlled directory, the user must supply a password. The Dir must be specified relative to the docroot. Multiple dir can be used. If no dir is set, the default value, "/", will be used.

realm = Realm
In the directory defined here, the WWW-Authenticate Realm is set to this value.

authmod = AuthMod
If an auth module is defined then AuthMod:auth(Arg, Auth) will be called for all access to the directory. The auth/2 function should return one of: true, false, {false, Realm}, {appmod, Mod}. If {appmod, Mod} is returned then a call to Mod:out401(Arg, Auth, Realm) will be used to deliver the content. If errormod_401 is defined, the call to Mod will be ignored. (Mod:out(Arg) is deprecated).

This can, for example, be used to implement cookie authentication. The auth() callback would check if a valid cookie header is present, if not it would return {appmod, ?MODULE} and the out401/1 function in the same module would return {redirect_local, "/login.html"}.

user = User:Password | "User:{Algo}Hash" | "User:{Algo}$Salt$Hash"
Inside this directory, the user User has access if the user supplies the password Password in the popup dialogue presented by the browser. It is also possible to provide a hashed password, encoded in base64. In that case, the algorithm used to hash the password must be set. Algo must be one of the following algorithms:


md5 | ripemd160 | sha | sha224 | sha256 | sha384 | sha512

It is possible to use salted hashes. If so, the Salt must be provided, encoded in base64. We can specify multiple users inside a single <auth> </auth> pair.

pam service = pam-service
If the item pam is part of the auth structure, Yaws will also try to authenticate the user using "pam" using the pam service indicated. Usual services are typically found under /etc/pam.d. Usual values are "system-auth" etc.

pam authentication is performed by an Erlang port program which is typically installed as suid root by the Yaws install script.

allow = all | ListOfHost
The allow directive affects which hosts can access an area of the server. Access can be controlled by IP address or IP address range. If all is specified, then all hosts are allowed access, subject to the configuration of the deny and order directives. To allow only particular hosts or groups of hosts to access the server, the host can be specified in any of the following formats:

A full IP address


allow = 10.1.2.3
allow = 192.168.1.104, 192.168.1.205

A network/netmask pair


allow = 10.1.0.0/255.255.0.0

A network/nnn CIDR specification


allow = 10.1.0.0/16

deny = all | ListOfHost
This directive allows access to the server to be restricted based on IP address. The arguments for the deny directive are identical to the arguments for the allow directive.

order = Ordering
The order directive, along with allow and deny directives, controls a three-pass access control system. The first pass processes either all allow or all deny directives, as specified by the order directive. The second pass parses the rest of the directives (deny or allow). The third pass applies to all requests which do not match either of the first two.

Ordering is one of (Default value is deny,allow):

First, all allow directives are evaluated; at least one must match, or the request is rejected. Next, deny directives are evaluated. If any matches, the request is rejected. Last, any requests which do not match an allow or a deny directive are denied by default.

First, all deny directives are evaluated; if any matched, the request is denied unless it also matches an allow directive. Any requests which do not match any allow or deny directives are permitted.

<opaque> ... </opaque>
This begins and ends an opaque configuration context for this server, where 'Key = Value' directives can be specified. These directives are ignored by Yaws (hence the name opaque), but can be accessed as a list of tuples {Key,Value} stored in the #sconf.opaque record entry. See also the description of the start_mod directive.

This mechanism can be used to pass data from a surrounding application into the individual .yaws pages.

Change the behavior of the {bindings, [...]} directive to treat all undefined keys found in returned out/1 content as if they were defined with an empty value, resulting in all undefined bindings effectively being stripped out of returned content. By default, strip_undefined_bindings is false, which means undefined bindings are ignored and their text is left as is in returned content.

This setting applies only for out/1 content, not to static pages or other returned content.

<extra_response_headers> ... </extra_response_headers>
This begins and ends a configuration context for extra response headers for this server, where directives for adding headers, erasing headers, and modules for handling extra response headers can be specified as follows:

Add Hdr with value Value to the response, but only if the response status code is one of these values:

200 OK
201 Created
204 No Content
206 Partial Content
301 Moved Permanently
302 Found
303 See Other
304 Not Modified
307 Temporary Redirect
308 Permanent Redirect

For any other status code, Hdr is not added.

Unconditionally add Hdr with value Value to the response, regardless of the response status code.

Remove Hdr and its associated value from the response.

Specifies a module to call to process extra response headers. Yaws calls Module:extra_response_headers/3 passing the following arguments:

An Erlang map holding the response headers with header name strings as keys and strings as header values

An #arg{} record representing the request. In cases where an extramod module is called following the invocation of an appmod, the #arg{} record field appmod_name indicates the name of the appmod that serviced the request, allowing the extramod to return extra HTTP headers appropriate for that appmod.

{StatusCode,Version}
A tuple where StatusCode is the numeric HTTP status code for the response, and Version is a tuple specifying the HTTP version, e.g. {1,1} for HTTP 1.1.

The Module:extra_response_headers/3 function should return either the original header map or a modified map where headers have been added, changed, or deleted. Added headers are not subject to the status code restrictions for the add extra response header directive, but the function can call yaws_api:http_extra_response_headers_add_status_codes/0 to retrieve the list of the status codes for which adding headers is normally allowed.

For response headers that can have multiple settings, such as Set-Cookie, multiple values can be specified in the extra response header map by using a value of {multi, [Value]} where [Value] is a list of one or more header values. The Set-Cookie header is a standard special case for which Yaws converts a multi header into a separate Set-Cookie header for each value; for other headers, Yaws converts a multi header into a single HTTP header with a comma-separated value.

Note that extra response headers do not apply to responses returned directly by any DispatchModule.

Setting options_asterisk_methods to a comma-separated list of HTTP Methods makes Yaws respond to an OPTIONS request that specifies a literal * as the target with a 200 OK status and an Allow header listing the specified Methods. If the configuration does not explicitly specify options_asterisk_methods, Yaws defaults to responding to OPTIONS * requests with a 200 OK status and an Allow header listing these HTTP methods:


GET, HEAD, POST, PUT, DELETE, OPTIONS

RFC 7231 section 4.3 lists the standard HTTP method names:


GET, HEAD, POST, PUT, DELETE, CONNECT, OPTIONS, TRACE

An options_asterisk_methods setting can include any of these HTTP method names as well as PATCH. PATCH is not mentioned in RFC 7231 but Yaws supports it. Yaws does not implement CONNECT, but it supports it in options_asterisk_methods because it's possible to implement support for it using a dispatchmod.

If options_asterisk_methods is set to an empty value, Yaws responds to OPTIONS * requests with status 400 Bad Request.

EXAMPLES

The following example defines a single server on port 80.


logdir = /var/log/yaws
<server www.mydomain.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www
</server>

And this example shows a similar setup but two web servers on the same IP address.


logdir = /var/log/yaws
<server www.mydomain.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www
</server>
<server www.funky.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www_funky_org
</server>

An example with www-authenticate and no access logging at all.


logdir = /var/log/yaws
<server www.mydomain.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www
access_log = false
<auth>
dir = secret/dir1
realm = foobar
user = jonny:verysecretpwd
user = benny:thequestion
user = ronny:havinganamethatendswithy
</auth>
</server>

An example specifying a user defined module to be called at startup, as well as some user specific configuration.


<server www.funky.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www_funky_org
start_mod = btt
<opaque>
mydbdir = /tmp
mylogdir = /tmp/log
</opaque>
</server>

An example specifying the GSSAPI/SPNEGO module (authmod_gssapi) to be used for authentication. This module requires egssapi version 0.1~pre2 or later available at http://www.hem.za.org/egssapi/.

The Kerberos5 keytab is specified as 'keytab = File' directive in opaque. This keytab should contain the keys of the HTTP service principal, 'HTTP/www.funky.org' in this example.


<server www.funky.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www_funky_org
start_mod = authmod_gssapi
<auth>
authmod = authmod_gssapi
dir = secret/dir1
</auth>
<opaque>
keytab = /etc/yaws/http.keytab
</opaque>
</server>

And finally a slightly more complex example with two servers on the same IP, and one SSL server on a different IP.

When there are more than one server on the same IP, and they have different names the server must be able to choose one of them if the client doesn't send a Host: header. Yaws will choose the first one defined in the conf file.


logdir = /var/log/yaws
max_num_cached_files = 8000
max_num_cached_bytes = 6000000
<server www.mydomain.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www
</server>
<server www.funky.org>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www_funky_org
</server>
<server www.funky.org>
port = 443
listen = 192.168.128.32
docroot = /var/yaws/www_funky_org
<ssl>
keyfile = /etc/funky.key
certfile = /etc/funky.cert
password = gazonk
</ssl>
</server>

Finally an example with virtual directories, vdirs.


<server server.domain>
port = 80
listen = 192.168.128.31
docroot = /var/yaws/www
arg_rewrite_mod = yaws_vdir
<opaque>
vdir = "/virtual1/ /usr/local/somewhere/notrelated/to/main/docroot"
vdir = "/myapp/ /some/other/path can include/spaces"
vdir = "/icons/ /usr/local/www/yaws/icons"
</opaque>
</server>

The first defined vdir can then be accessed at or under http://server.domain/virtual1/ or http://server.domain/virtual1

AUTHOR

Written by Claes Wikstrom

SEE ALSO

yaws(1) erl(1)