librrd contains most of the functionality in RRDTool. The command line utilities and language bindings are often just wrappers around the code contained in librrd.

This manual page documents the librrd API.

NOTE: This document is a work in progress, and should be considered incomplete as long as this warning persists. For more information about the librrd functions, always consult the source code.


rrd_dump_cb_r(char *filename, int opt_header, rrd_output_callback_t cb, void *user)

In some situations it is necessary to get the output of rrd_dump without writing it to a file or the standard output. In such cases an application can ask rrd_dump_cb_r to call an user-defined function each time there is output to be stored somewhere. This can be used, to e.g. directly feed an XML parser with the dumped output or transfer the resulting string in memory.

The arguments for rrd_dump_cb_r are the same as for rrd_dump_opt_r except that the output filename parameter is replaced by the user-defined callback function and an additional parameter for the callback function that is passed untouched, i.e. to store information about the callback state needed for the user-defined callback to function properly.

Recent versions of rrd_dump_opt_r internally use this callback mechanism to write their output to the file provided by the user.

    size_t rrd_dump_opt_cb_fileout(
        const void *data,
        size_t len,
        void *user)
        return fwrite(data, 1, len, (FILE *)user);

The associated call for rrd_dump_cb_r looks like

    res = rrd_dump_cb_r(filename, opt_header,
        rrd_dump_opt_cb_fileout, (void *)out_file);

where the last parameter specifies the file handle rrd_dump_opt_cb_fileout should write to. There's no specific condition for the callback to detect when it is called for the first time, nor for the last time. If you require this for initialization and cleanup you should do those tasks before and after calling rrd_dump_cr_r respectively.

rrd_fetch_cb_register(rrd_fetch_cb_t c)

If your data does not reside in rrd files, but you would like to draw charts using the rrd_graph functionality, you can supply your own rrd_fetch function and register it using the rrd_fetch_cb_register function.

The argument signature and api must be the same of the callback function must be aequivalent to the on of rrd_fetch_fn in rrd_fetch.c.

To activate the callback function you can use the pseudo filename cb//free_form_text.

Note that rrdtool graph will not ask the same rrd for data twice. It determines this by building a key out of the values supplied to the fetch function. If the values are the same, the previous answer will be used.



Generates random numbers just like random(). This further ensures that the random number generator is seeded exactly once per process.


an rrd aware string to double converter which sets rrd_error in if there is a problem and uses the return code exclusively for conversion status reporting.


works like normal strtod, but it is locale independent (and thread safe)


works like normal snprintf but it is locale independent (and thread safe)

rrd_add_ptr(void ***dest, size_t *dest_size, void *src)

Dynamically resize the array pointed to by dest. dest_size is a pointer to the current size of dest. Upon successful realloc(), the dest_size is incremented by 1 and the src pointer is stored at the end of the new dest. Returns 1 on success, 0 on failure.

    type **arr = NULL;
    type *elem = "whatever";
    size_t arr_size = 0;
    if (!rrd_add_ptr(&arr, &arr_size, elem))
rrd_add_ptr_chunk(void ***dest, size_t *dest_size, void *src, size_t *alloc, size_t chunk)

Like rrd_add_ptr, except the destination is allocated in chunks of chunk. alloc points to the number of entries allocated, whereas dest_size points to the number of valid pointers. If more pointers are needed, chunk pointers are allocated and alloc is increased accordingly. alloc must be >= dest_size.

This method improves performance on hosts with expensive realloc().

rrd_add_strdup(char ***dest, size_t *dest_size, char *src)

Like rrd_add_ptr, except adds a strdup of the source string.

    char **arr = NULL;
    size_t arr_size = NULL;
    char *str  = "example text";
    if (!rrd_add_strdup(&arr, &arr_size, str))
rrd_add_strdup_chunk(char ***dest, size_t *dest_size, char *src, size_t *alloc, size_t chunk)

Like rrd_add_strdup, except the destination is allocated in chunks of chunk. alloc points to the number of entries allocated, whereas dest_size points to the number of valid pointers. If more pointers are needed, chunk pointers are allocated and alloc is increased accordingly. alloc must be >= dest_size.

rrd_free_ptrs(void ***src, size_t *cnt)

Free an array of pointers allocated by rrd_add_ptr or rrd_add_strdup. Also frees the array pointer itself. On return, the source pointer will be NULL and the count will be zero.

    /* created as above */
    rrd_free_ptrs(&arr, &arr_size);
    /* here, arr == NULL && arr_size == 0 */
rrd_mkdir_p(const char *pathname, mode_t mode)

Create the directory named pathname including all of its parent directories (similar to mkdir -p on the command line - see mkdir(1) for more information). The argument mode specifies the permissions to use. It is modified by the process's umask. See mkdir(2) for more details.

The function returns 0 on success, a negative value else. In case of an error, errno is set accordingly. Aside from the errors documented in mkdir(2), the function may fail with the following errors:


pathname is NULL or the empty string.


Insufficient memory was available.

any error returned by stat(2)

In contrast to mkdir(2), the function does not fail if pathname already exists and is a directory.

rrd_scaled_duration (const char * token, unsigned long divisor, unsigned long * valuep)

Parse a token in a context where it contains a count (of seconds or PDP instances), or a duration that can be converted to a count by representing the duration in seconds and dividing by some scaling factor. For example, if a user would natively express a 3 day archive of samples collected every 2 minutes, the sample interval can be represented by 2m instead of 120, and the archive duration by 3d (to be divided by 120) instead of 2160 (3*24*60*60 / 120). See more examples in "STEP, HEARTBEAT, and Rows As Durations" in rrdcreate.

token must be a number with an optional single-character suffix encoding the scaling factor:


indicates seconds


indicates minutes. The value is multipled by 60.


indicates hours. The value is multipled by 3600 (or 60m).


indicates days. The value is multipled by 86400 (or 24h).


indicates weeks. The value is multipled by 604800 (or 7d).


indicates months. The value is multipled by 2678400 (or 31d). (Note this factor accommodates the maximum number of days in a month.)


indicates years. The value is multipled by 31622400 (or 366d). (Note this factor accommodates leap years.)

divisor is a positive value representing the duration in seconds of an interval that the desired result counts.

valuep is a pointer to where the decoded value will be stored if the conversion is successful.

The initial characters of token must be the base-10 representation of a positive integer, or the conversion fails.

If the remainder token is empty (no suffix), it is a count and no scaling is performed.

If token has one of the suffixes above, the count is multipled to convert it to a duration in seconds. The resulting number of seconds is divided by divisor to produce a count of intervals each of duration divisor seconds. If division would produce a remainder (e.g., 5m (300 seconds) divided by 90s), the conversion is invalid.

If token has unrecognized trailing characters the conversion fails.

The function returns a null pointer if the conversion was successful and valuep has been updated to the scaled value. On failure, it returns a text diagnostic suitable for use in user error messages.


RRD Contributors <>

NOTE: The content of this website is accessible with any browser. The graphical design though relies completely on CSS2 styles. If you see this text, this means that your browser does not support CSS2. Consider upgrading to a standard conformant browser like Mozilla Firefox or Opera but also Apple's Safari or KDE's Konqueror for example. It may also be that you are looking at a mirror page which did not copy the CSS for this page. Or if some pictu res are missing, then the mirror may not have picked up the contents of the inc directory.