1RRDCACHED(1) rrdtool RRDCACHED(1)
2
6 rrdcached - Data caching daemon for rrdtool
7
9 rrdcached [-P permissions] [-l address] [-s group] [-w timeout]
10 [-z delay] [-f timeout] [-p pid_file] [-t write_threads]
11 [-j journal_dir] [-F] [-g] [-b base_dir [-B]]
12
14 rrdcached is a daemon that receives updates to existing RRD files,
15 accumulates them and, if enough have been received or a defined time
16 has passed, writes the updates to the RRD file. A flush command may be
17 used to force writing of values to disk, so that graphing facilities
18 and similar can work with up-to-date data.
19 The daemon was written with big setups in mind. Those setups usually
21 run into IO related problems sooner or later for reasons that are
22 beyond the scope of this document. Check the wiki at the RRDtool
23 homepage for details. Also check "SECURITY CONSIDERATIONS" below before
24 using this daemon! A detailed description of how the daemon operates
25 can be found in the "HOW IT WORKS" section below.
26
28 -l address
29 Tells the daemon to bind to address and accept incoming connections
30 on that socket. If address begins with "unix:", everything
31 following that prefix is interpreted as the path to a UNIX domain
32 socket. Otherwise the address or node name are resolved using
33 "getaddrinfo()".
34 For network sockets, a port may be specified by using the form
36 "[address]:port". If the address is an IPv4 address or a fully
37 qualified domain name (i. e. the address contains at least one dot
38 (".")), the square brackets can be omitted, resulting in the
39 (simpler) "address:port" pattern. The default port is 42217/tcp. If
40 you specify a network socket, it is mandatory to read the "SECURITY
41 CONSIDERATIONS" section.
42 The following formats are accepted. Please note that the address of
44 the UNIX domain socket must start with a slash in the second case!
45 unix:</path/to/unix.sock>
47 /<path/to/unix.sock>
48 <hostname-or-ip>
49 [<hostname-or-ip>]:<port>
50 <hostname-or-ipv4>:<port>
51 If the -l option is not specified the default address,
53 "unix:/tmp/rrdcached.sock", will be used.
54 -s group_name|gid
56 Set the group permissions of a UNIX domain socket. The option
57 accepts either a numeric group id or group name. That group will
58 then have both read and write permissions (the socket will have
59 file permissions 0750) for the socket and, therefore, is able to
60 send commands to the daemon. This may be useful in cases where you
61 cannot easily run all RRD processes with the same user privileges
62 (e.g. graph generating CGI scripts that typically run in the
63 permission context of the web server).
64 This option affects the following UNIX socket addresses (the
66 following -l options) or the default socket (if no -l options have
67 been specified), i.e., you may specify different settings for
68 different sockets.
69 The default is not to change ownership or permissions of the socket
71 and, thus, use the system default.
72 -m mode
74 Set the file permissions of a UNIX domain socket. The option
75 accepts an octal number representing the bit pattern for the mode
76 (see chmod(1) for details).
77 Please note that not all systems honor this setting. On Linux,
79 read/write permissions are required to connect to a UNIX socket.
80 However, many BSD-derived systems ignore permissions for UNIX
81 sockets. See unix(7) for details.
82 This option affects the following UNIX socket addresses (the
84 following -l options) or the default socket (if no -l options have
85 been specified), i.e., you may specify different settings for
86 different sockets.
87 The default is not to change ownership or permissions of the socket
89 and, thus, use the system default.
90 -P command[,command[,...]]
92 Specifies the commands accepted via a network socket. This allows
93 administrators of RRDCacheD to control the actions accepted from
94 various sources.
95 The arguments given to the -P option is a comma separated list of
97 commands. For example, to allow the "FLUSH" and "PENDING" commands
98 one could specify:
99 rrdcached -P FLUSH,PENDING $MORE_ARGUMENTS
101 The -P option affects the following socket addresses (the following
103 -l options) or the default socket (if no -l options have been
104 specified). In the following example, only the IPv4 network socket
105 (address 10.0.0.1) will be restricted to the "FLUSH" and "PENDING"
106 commands:
107 rrdcached -l unix:/some/path -P FLUSH,PENDING -l 10.0.0.1
109 A complete list of available commands can be found in the section
111 "Valid Commands" below. There are two minor special exceptions:
112 · The "HELP" and "QUIT" commands are always allowed.
114 · If the "BATCH" command is accepted, the . command will
116 automatically be accepted, too.
117 Please also read "SECURITY CONSIDERATIONS" below.
119 -w timeout
121 Data is written to disk every timeout seconds. If this option is
122 not specified the default interval of 300 seconds will be used.
123 -z delay
125 If specified, rrdcached will delay writing of each RRD for a random
126 number of seconds in the range [0,delay). This will avoid too many
127 writes being queued simultaneously. This value should be no
128 greater than the value specified in -w. By default, there is no
129 delay.
130 -f timeout
132 Every timeout seconds the entire cache is searched for old values
133 which are written to disk. This only concerns files to which
134 updates have stopped, so setting this to a high value, such as
135 3600 seconds, is acceptable in most cases. This timeout defaults to
136 3600 seconds.
137 -p file
139 Sets the name and location of the PID-file. If not specified, the
140 default, "$localststedir/run/rrdcached.pid" will be used.
141 -t write_threads
143 Specifies the number of threads used for writing RRD files. The
144 default is 4. Increasing this number will allow rrdcached to have
145 more simultaneous I/O requests into the kernel. This may allow the
146 kernel to re-order disk writes, resulting in better disk
147 throughput.
148 -j dir
150 Write updates to a journal in dir. In the event of a program or
151 system crash, this will allow the daemon to write any updates that
152 were pending at the time of the crash.
153 On startup, the daemon will check for journal files in this
155 directory. If found, all updates therein will be read into memory
156 before the daemon starts accepting new connections.
157 The journal will be rotated with the same frequency as the flush
159 timer given by -f.
160 When journaling is enabled, the daemon will use a fast shutdown
162 procedure. Rather than flushing all files to disk, it will make
163 sure the journal is properly written and exit immediately.
164 Although the RRD data files are not fully up-to-date, no
165 information is lost; all pending updates will be replayed from the
166 journal next time the daemon starts up.
167 To disable fast shutdown, use the -F option.
169 -F ALWAYS flush all updates to the RRD data files when the daemon is
171 shut down, regardless of journal setting.
172 -g Run in the foreground. The daemon will not fork().
174 -b dir
176 The daemon will change into a specific directory at startup. All
177 files passed to the daemon, that are specified by a relative path,
178 will be interpreted to be relative to this directory. If not given
179 the default, "/tmp", will be used.
180 +------------------------+------------------------+
182 ! Command line ! File updated !
183 +------------------------+------------------------+
184 ! foo.rrd ! /tmp/foo.rrd !
185 ! foo/bar.rrd ! /tmp/foo/bar.rrd !
186 ! /var/lib/rrd/foo.rrd ! /var/lib/rrd/foo.rrd !
187 +------------------------+------------------------+
188 Paths given on the command line and paths actually
189 updated by the daemon, assuming the base directory
190 "/tmp".
191 WARNING: The paths up to and including the base directory MUST NOT
193 BE symbolic links. In other words, if the base directory is
194 specified as:
195 -b /base/dir/somewhere
197 ... then NONE of the following should be symbolic links:
199 /base
201 /base/dir
202 /base/dir/somewhere
203 -B Only permit writes into the base directory specified in -b (and any
205 sub-directories). This does NOT detect symbolic links. Paths
206 containing "../" will also be blocked.
207
209 The following commands may be made aware of the rrdcached using the
210 command line argument --daemon or the environment variable
211 RRDCACHED_ADDRESS:
212 · dump
214 · fetch
216 · flush
218 · graph
220 · graphv
222 · info
224 · last
226 · lastupdate
228 · update
230 · xport
232 The update command can send values to the daemon instead of writing
234 them to the disk itself. All other commands can send a FLUSH command
235 (see below) to the daemon before accessing the files, so they work with
236 up-to-date data even if the cache timeout is large.
237
239 The daemon reports errors in one of two ways: During startup, error
240 messages are printed to "STDERR". One of the steps when starting up is
241 to fork to the background and closing "STDERR" - after this writing
242 directly to the user is no longer possible. Once this has happened, the
243 daemon will send log messages to the system logging daemon using
244 syslog(3). The facility used is "LOG_DAEMON".
245
247 When receiving an update, rrdcached does not write to disk but looks
248 for an entry for that file in its internal tree. If not found, an entry
249 is created including the current time (called "First" in the diagram
250 below). This time is not the time specified on the command line but the
251 time the operating system considers to be "now". The value and time of
252 the value (called "Time" in the diagram below) are appended to the tree
253 node.
254 When appending a value to a tree node, it is checked whether it's time
256 to write the values to disk. Values are written to disk if
257 "now() - First >= timeout", where "timeout" is the timeout specified
258 using the -w option, see "OPTIONS". If the values are "old enough" they
259 will be enqueued in the "update queue", i. e. they will be appended to
260 the linked list shown below. Because the tree nodes and the elements
261 of the linked list are the same data structures in memory, any update
262 to a file that has already been enqueued will be written with the next
263 write to the RRD file, too.
264 A separate "update thread" constantly dequeues the first element in the
266 update queue and writes all its values to the appropriate file. So as
267 long as the update queue is not empty files are written at the highest
268 possible rate.
269 Since the timeout of files is checked only when new values are added to
271 the file, "dead" files, i. e. files that are not updated anymore, would
272 never be written to disk. Therefore, every now and then, controlled by
273 the -f option, the entire tree is walked and all "old" values are
274 enqueued. Since this only affects "dead" files and walking the tree is
275 relatively expensive, you should set the "flush interval" to a
276 reasonably high value. The default is 3600 seconds (one hour).
277 The downside of caching values is that they won't show up in graphs
279 generated from the RRD files. To get around this, the daemon provides
280 the "flush command" to flush specific files. This means that the file
281 is inserted at the head of the update queue or moved there if it is
282 already enqueued. The flush command will return only after the file's
283 pending updates have been written to disk.
284 +------+ +------+ +------+
286 ! head ! ! root ! ! tail !
287 +---+--+ +---+--+ +---+--+
288 ! /\ !
289 ! / \ !
290 ! /\ /\ !
291 ! /\/\ \ `----------------- ... --------, !
292 V / `-------, ! V
293 +---+----+---+ +------+-----+ +---+----+---+
294 ! File: foo ! ! File: bar ! ! File: qux !
295 ! First: 101 ! ! First: 119 ! ! First: 180 !
296 ! Next:&bar -+--->! Next:&... -+---> ... --->! Next:NULL !
297 | Prev:NULL !<---+-Prev:&foo !<--- ... ----+-Prev: &... !
298 +============+ +============+ +============+
299 ! Time: 100 ! ! Time: 120 ! ! Time: 180 !
300 ! Value: 10 ! ! Value: 0.1 ! ! Value: 2,2 !
301 +------------+ +------------+ +------------+
302 ! Time: 110 ! ! Time: 130 ! ! Time: 190 !
303 ! Value: 26 ! ! Value: 0.1 ! ! Value: 7,3 !
304 +------------+ +------------+ +------------+
305 : : : : : :
306 +------------+ +------------+ +------------+
307 ! Time: 230 ! ! Time: 250 ! ! Time: 310 !
308 ! Value: 42 ! ! Value: 0.2 ! ! Value: 1,2 !
309 +------------+ +------------+ +------------+
310 The above diagram demonstrates:
312 · Files/values are stored in a (balanced) tree.
314 · Tree nodes and entries in the update queue are the same data
316 structure.
317 · The local time ("First") and the time specified in updates ("Time")
319 may differ.
320 · Timed out values are inserted at the "tail".
322 · Explicitly flushed values are inserted at the "head".
324 · ASCII art rocks.
326
328 Authentication
329 If your rrdtool installation was built without libwrap there is no form
330 of authentication for clients connecting to the rrdcache daemon!
331 If your rrdtool installation was built with libwrap then you can use
333 hosts_access to restrict client access to the rrdcache daemon
334 (rrdcached). For more information on how to use hosts_access to
335 restrict access to the rrdcache daemon you should read the
336 hosts_access(5) man pages.
337 It is still highly recommended to install a packet filter or similar
339 mechanism to prevent unauthorized connections. Unless you have a
340 dedicated VLAN or VPN for this, using network sockets is probably a bad
341 idea!
342 Authorization
344 There is minimal per-socket authorization.
345 Authorization is currently done on a per-socket basis. That means each
347 socket has a list of commands it will accept and it will accept. It
348 will accept only those commands explicitly listed but it will
349 (currently) accept these commands from anyone reaching the socket.
350 If the networking sockets are to be used, it is necessary to restrict
352 the accepted commands to those needed by external clients. If, for
353 example, external clients want to draw graphs of the cached data, they
354 should only be allowed to use the "FLUSH" command.
355 Encryption
357 There is no encryption.
358 Again, this may be added in the future, but for the time being it is
360 your job to keep your private data private. Install a VPN or an
361 encrypted tunnel if you statistics are confidential!
362 Sanity checking
364 There is no sanity checking.
365 The daemon will blindly write to any file it gets told, so you really
367 should create a separate user just for this daemon. Also it does not do
368 any sanity checks, so if it gets told to write values for a time far in
369 the future, your files will be messed up good!
370 Conclusion
372 · Security is the job of the administrator.
373 · We recommend to allow write access via UNIX domain sockets only.
375 · You have been warned.
377
379 The daemon communicates with clients using a line based ASCII protocol
380 which is easy to read and easy to type. This makes it easy for scripts
381 to implement the protocol and possible for users to use telnet to
382 connect to the daemon and test stuff "by hand".
383 The protocol is line based, this means that each record consists of one
385 or more lines. A line is terminated by the line feed character 0x0A,
386 commonly written as "\n". In the examples below, this character will be
387 written as "<LF>" ("line feed").
388 After the connection has been established, the client is expected to
390 send a "command". A command consists of the command keyword, possibly
391 some arguments, and a terminating newline character. For a list of
392 commands, see "Valid Commands" below.
393 Example:
395 FLUSH /tmp/foo.rrd<LF>
397 The daemon answers with a line consisting of a status code and a short
399 status message, separated by one or more space characters. A negative
400 status code signals an error, a positive status code or zero signal
401 success. If the status code is greater than zero, it indicates the
402 number of lines that follow the status line.
403 Examples:
405 0 Success<LF>
407 2 Two lines follow<LF>
409 This is the first line<LF>
410 And this is the second line<LF>
411 Valid Commands
413 The following commands are understood by the daemon:
414 FLUSH filename
416 Causes the daemon to put filename to the head of the update queue
417 (possibly moving it there if the node is already enqueued). The
418 answer will be sent after the node has been dequeued.
419 FLUSHALL
421 Causes the daemon to start flushing ALL pending values to disk.
422 This returns immediately, even though the writes may take a long
423 time.
424 PENDING filename
426 Shows any "pending" updates for a file, in order. The updates
427 shown have not yet been written to the underlying RRD file.
428 FORGET filename
430 Removes filename from the cache. Any pending updates WILL BE LOST.
431 QUEUE
433 Shows the files that are on the output queue. Returns zero or more
434 lines in the following format, where <num_vals> is the number of
435 values to be written for the <file>:
436 <num_vals> <file>
438 HELP [command]
440 Returns a short usage message. If no command is given, or command
441 is HELP, a list of commands supported by the daemon is returned.
442 Otherwise a short description, possibly containing a pointer to a
443 manual page, is returned. Obviously, this is meant for interactive
444 usage and the format in which the commands and usage summaries are
445 returned is not well defined.
446 STATS
448 Returns a list of metrics which can be used to measure the daemons
449 performance and check its status. For a description of the values
450 returned, see "Performance Values" below.
451 The format in which the values are returned is similar to many
453 other line based protocols: Each value is printed on a separate
454 line, each consisting of the name of the value, a colon, one or
455 more spaces and the actual value.
456 Example:
458 9 Statistics follow
460 QueueLength: 0
461 UpdatesReceived: 30
462 FlushesReceived: 2
463 UpdatesWritten: 13
464 DataSetsWritten: 390
465 TreeNodesNumber: 13
466 TreeDepth: 4
467 JournalBytes: 190
468 JournalRotate: 0
469 UPDATE filename values [values ...]
471 Adds more data to a filename. This is the operation the daemon was
472 designed for, so describing the mechanism again is unnecessary.
473 Read "HOW IT WORKS" above for a detailed explanation.
474 Note that rrdcached only accepts absolute timestamps in the update
476 values. Updates strings like "N:1:2:3" are automatically converted
477 to absolute time by the RRD client library before sending to
478 rrdcached.
479 WROTE filename
481 This command is written to the journal after a file is successfully
482 written out to disk. It is used during journal replay to determine
483 which updates have already been applied. It is only valid in the
484 journal; it is not accepted from the other command channels.
485 BATCH
487 This command initiates the bulk load of multiple commands. This is
488 designed for installations with extremely high update rates, since
489 it permits more than one command to be issued per read() and
490 write().
491 All commands are executed just as they would be if given
493 individually, except for output to the user. Messages indicating
494 success are suppressed, and error messages are delayed until the
495 client is finished.
496 Command processing is finished when the client sends a dot (".") on
498 its own line. After the client has finished, the server responds
499 with an error count and the list of error messages (if any). Each
500 error messages indicates the number of the command to which it
501 corresponds, and the error message itself. The first user command
502 after BATCH is command number one.
503 client: BATCH
505 server: 0 Go ahead. End with dot '.' on its own line.
506 client: UPDATE x.rrd 1223661439:1:2:3 <--- command #1
507 client: UPDATE y.rrd 1223661440:3:4:5 <--- command #2
508 client: and so on...
509 client: .
510 server: 2 Errors
511 server: 1 message for command 1
512 server: 12 message for command 12
513 QUIT
515 Disconnect from rrdcached.
516 Performance Values
518 The following counters are returned by the STATS command:
519 QueueLength (unsigned 64bit integer)
521 Number of nodes currently enqueued in the update queue.
522 UpdatesReceived (unsigned 64bit integer)
524 Number of UPDATE commands received.
525 FlushesReceived (unsigned 64bit integer)
527 Number of FLUSH commands received.
528 UpdatesWritten (unsigned 64bit integer)
530 Total number of updates, i. e. calls to "rrd_update_r", since the
531 daemon was started.
532 DataSetsWritten (unsigned 64bit integer)
534 Total number of "data sets" written to disk since the daemon was
535 started. A data set is one or more values passed to the UPDATE
536 command. For example: "1223661439:123:456" is one data set with two
537 values. The term "data set" is used to prevent confusion whether
538 individual values or groups of values are counted.
539 TreeNodesNumber (unsigned 64bit integer)
541 Number of nodes in the cache.
542 TreeDepth (unsigned 64bit integer)
544 Depth of the tree used for fast key lookup.
545 JournalBytes (unsigned 64bit integer)
547 Total number of bytes written to the journal since startup.
548 JournalRotate (unsigned 64bit integer)
550 Number of times the journal has been rotated since startup.
551
553 SIGINT and SIGTERM
554 The daemon exits normally on receipt of either of these signals.
555 Pending updates are handled in accordance with the -j and -F
556 options.
557 SIGUSR1
559 The daemon exits AFTER flushing all updates out to disk. This may
560 take a while.
561 SIGUSR2
563 The daemon exits immediately, without flushing updates out to disk.
564 Pending updates will be replayed from the journal when the daemon
565 starts up again. WARNING: if journaling (-j) is NOT enabled, any
566 pending updates WILL BE LOST.
567
569 No known bugs at the moment.
570
572 rrdtool, rrdgraph
573
575 Florian Forster <octo at verplant.org>
576 Both rrdcached and this manual page have been written by Florian.
578
580 kevin brintnall <kbrint@rufus.net>
5811.4.8 2013-05-23 RRDCACHED(1)