Configuration Variable Implementation¶
Adding a new configuration variable in
records.config requires a number
of steps which are mostly documented here.
Before adding a new configuration variable, please discuss it on the mailing list. It will commonly be the case that a better name, or a more general approach to the problem which solves several different issues, may be suggested.
Defining the Variable¶
To begin, the new configuration variables must be added to
This contains a long array of configuration variable records. The fields for
each record are:
Type of record. The valid values are:
- Undefined record.
- General configuration variable.
- Process related statistic.
- Local statistic.
- Plugin created statistic.
RECT_CONFIGshould be used.
- The fully qualified name of the configuration variable. Although there appears to be a hierarchial naming scheme, that’s just a convention, and it is not actually used by the code. Nonetheless, new variables should adhere to the hierarchial scheme.
- The data type of the value. It should be one of
- The default value for the variable. This is always a string regardless of the value_type.
Information about how the variable is updated. The valid values are:
- Behavior is unknown or unspecified.
- This can be updated via command line tools.
- The traffic_server process must be restarted for a new value to take effect.
- The traffic_manager process must be restarted for a new value to take effect.
- Effectively a boolean that specifies if the record is required to be present,
RR_NULLmeaning not required and
RR_REQUIREDindicating that it is required. Given that using
RR_REQUIREDwould be a major incompatibility,
RR_NULLis generally the better choice.
Additional type checking. It is unclear if this is actually implemented. The valid values are:
- No additional checking.
- Verify the value is a string.
- Verify the value is an integer.
- Verify the value is an IP address. Unknown if this checks for IPv6.
- This provides a regular expressions (PCRE format) for validating the value,
beyond the basic type validation performed by
RecCheckT. This can be
NULLif there is no regular expression to use.
Access control. The valid values are:
- The value is read / write.
- The value is read only.
- No access to the value; only privileged level parts of ATS can access the value.
The primary effort in defining a configuration variable is handling updates,
traffic_ctl config reload. This is handled in a generic way, as
described in the next section, or in a more specialized way
(built on top of the generic mechanism) for HTTP related configuration
variables. This is only needed if the variable is marked as dynamically
RECU_DYNAMIC) although HTTP configuration variables should be
dynamic if possible.
Documentation and Defaults¶
A configuration variable should be documented in
are many examples in the file already that can be used for guidance. The general
format is to use the tag
The arguments to this are the same as for the configuration file. The documentation generator will pick out key bits and use them to decorate the entry. In particular if a value is present it will be removed and used as the default value. You can attach some additional options to the variable. These are:
- The variable can be reloaded via command line on a running Traffic Server.
- Specify the units for the value. This is critical for variables that use unexpected or non-obvious metrics, such as minutes instead of seconds, or disk sectors instead of bytes.
- Mark a variable as deprecated.
reloadable: units: minutes deprecated:
If you need to refer to another configuration variable in the documentation, you can use the form
This will display the name as a link to the full definition.
In general, a new configuration variable should not be present in the default
records.config. If it is added, such defaults should be added to the
proxy/config/records.config.default.in. This is used to generate the
records.config. Just add the variable to the file in an
appropriate place with a proper default as this will now override whatever
default you put in the code for new installs.
The simplest mechanism for handling updates is the
family of functions. This mechanism will cause the value in the indicated
instance to be updated in place when an update to
This is done asynchronously using atomic operations. Use of these variables must
keep that in mind.
If a variable requires additional handling when updated a callback can be
registered which is called when the variable is updated. This is what the
REC_EstablishStaticConfigXXX calls do internally with a callback that simply
reads the new value and writes it to storage indicated by the call parameters.
The functions used are the
link_XXX static functions in
To register a configuration variable callback, call
with the arguments:
- The variable name.
- A function with the signature
<int (char const* name, RecDataT type, RecData data, void* cookie)>. The name value passed is the same as the name passed to the registration function as is the cookie argument. The type and data are the new value for the variable. The return value is currently ignored. For future compatibility return
- A value passed to the callback. This is only for the callback, the internals simply store it and pass it on.
callback is called under lock so it should be quick and not block. If that is necessary a continuation should be scheduled to handle the required action.
The callback occurs asynchronously. For HTTP variables as described in the next section, this is handled by the more specialized HTTP update mechanisms. Otherwise it is the implementor’s responsibility to avoid race conditions.
HTTP Configuation Values¶
Variables used for HTTP processing should be declared as members of the
HTTPConfigParams structure (but see Overridable Variables for
further details) and use the specialized HTTP update mechanisms which handle
synchronization and initialization issues.
The configuration logic maintains two copies of the
structure, the master copy and the current copy. The master copy is kept in the
m_master member of the
HttpConfig singleton. The current copy is kept in
the ConfigProcessor. The goal is to provide a (somewhat) atomic update for
configuration variables which are loaded individually in to the master copy as
updates are received and then bulk copied to a new instance which is then
swapped in as the current copy. The HTTP state machine interacts with this
mechanism to avoid race conditions.
For each variable, a mapping between the variable name and the appropriate
member in the master copy should be established between in the
HttpEstablishStaticConfigXXX functions should be used unless
there is a strong, explicit reason to not do so.
HTTPConfig::reconfigure method handles the current copy of the HTTP
configuration variables. Logic should be added here to copy the value from the
master copy to the current copy. Generally this will be a simple assignment. If
there are dependencies between variables, those should be checked and enforced
in this method.
HTTP related variables that are changeable per transaction are stored in the
OverridableHttpConfigParams structure, an instance of which is the
HTTPConfigParams and therefore the points in the previous section
still apply. The only difference for that is the further
.oride member specifier in the structure references.
The variable is required to be accessible from the transaction API. In addition to any custom API functions used to access the value, the following items are required for generic access:
- Add a value to the
- Augment the
TSHttpTxnConfigFindfunction to return this enumeration value when given the name of the configuration variable. Be sure to count the charaters very carefully.
- Augment the
InkAPI.ccto return a pointer to the appropriate member of
OverridableHttpConfigParamsand set the type if not a byte value.
- Update the testing logic in
InkAPITest.ccby adding the string name of the configuration variable to the
- Update the Lua plugin enumeration
- Update the documentation of TSHttpOverridableConfig in
A relatively new feature for overridable variables is the ability to keep them in more natural data types and convert as needed to the API types. This in turns enables defining the configuration locally in a module and then “exporting” it to the API interface. Modules then do not have to include headers for all types in all overridable configurations.
The conversion is done through an instance of
MgmtConverter. This has 6 points to
conversions, a load and store function for each of the types
MgmtByte type is handled by the
MgmtInt conversions. In general
each overridable variable will specify two of these, a load and store for a specific type, although
it is possible to provide other pairs, e.g. if a value is an enumeration can should be settable
as a string as well as an integer.
The module is responsible for creating an instance of
MgmtConverter with the appropriate
load / store function pairs set. The declaration must be visible in the proxy/InkAPI.cc
file. The function
_conf_to_memberp sets up the conversion. For the value of the enumeration
TSOverridableConfigKey that specifies the overridable variable, code is added to specify
the member and the conversion. There are default converters for the API types and if the overridable
is one of those, it is only necessary to call
_memberp_to_generic passing in a pointer to
the variable. For a variable with conversion, ret should be set to point to the variable and
conv set to point to the converter for that variable. If multiple variables are of the same
type they can use the same converter because a pointer to the specific member is passed to the