2
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
3
<html xmlns="http://www.w3.org/1999/xhtml">
5
<title>rrdcached</title>
6
<meta http-equiv="content-type" content="text/html; charset=utf-8" />
7
<link rev="made" href="mailto:root@localhost" />
10
<body style="background-color: white">
15
<p><a name="__index__"></a></p>
20
<li><a href="#name">NAME</a></li>
21
<li><a href="#synopsis">SYNOPSIS</a></li>
22
<li><a href="#description">DESCRIPTION</a></li>
23
<li><a href="#options">OPTIONS</a></li>
24
<li><a href="#affected_rrdtool_commands">AFFECTED RRDTOOL COMMANDS</a></li>
25
<li><a href="#error_reporting">ERROR REPORTING</a></li>
26
<li><a href="#how_it_works">HOW IT WORKS</a></li>
27
<li><a href="#security_considerations">SECURITY CONSIDERATIONS</a></li>
30
<li><a href="#authentication">Authentication</a></li>
31
<li><a href="#authorization">Authorization</a></li>
32
<li><a href="#encryption">Encryption</a></li>
33
<li><a href="#sanity_checking">Sanity checking</a></li>
34
<li><a href="#conclusion">Conclusion</a></li>
37
<li><a href="#protocol">PROTOCOL</a></li>
40
<li><a href="#valid_commands">Valid Commands</a></li>
41
<li><a href="#performance_values">Performance Values</a></li>
44
<li><a href="#signals">SIGNALS</a></li>
45
<li><a href="#bugs">BUGS</a></li>
46
<li><a href="#see_also">SEE ALSO</a></li>
47
<li><a href="#author">AUTHOR</a></li>
48
<li><a href="#contributors">CONTRIBUTORS</a></li>
60
<h1><a name="name">NAME</a></h1>
61
<p>rrdcached - Data caching daemon for rrdtool</p>
65
<h1><a name="synopsis">SYNOPSIS</a></h1>
66
<p><strong>rrdcached</strong>
67
[<strong>-P</strong> <em>permissions</em>]
68
[<strong>-l</strong> <em>address</em>]
69
[<strong>-s</strong> <em>group</em>]
70
[<strong>-w</strong> <em>timeout</em>]
71
[<strong>-z</strong> <em>delay</em>]
72
[<strong>-f</strong> <em>timeout</em>]
73
[<strong>-p</strong> <em>pid_file</em>]
74
[<strong>-t</strong> <em>write_threads</em>]
75
[<strong>-j</strong> <em>journal_dir</em>]
78
[<strong>-b</strong> <em>base_dir</em> [<strong>-B</strong>]]</p>
82
<h1><a name="description">DESCRIPTION</a></h1>
83
<p><strong>rrdcached</strong> is a daemon that receives updates to existing RRD files,
84
accumulates them and, if enough have been received or a defined time has
85
passed, writes the updates to the RRD file. A <em>flush</em> command may be used to
86
force writing of values to disk, so that graphing facilities and similar can
87
work with up-to-date data.</p>
88
<p>The daemon was written with big setups in mind. Those setups usually run into
89
IO related problems sooner or later for reasons that are beyond the scope
90
of this document. Check the wiki at the RRDtool homepage for details. Also
91
check <a href="#security_considerations">SECURITY CONSIDERATIONS</a> below before using this daemon! A detailed
92
description of how the daemon operates can be found in the <a href="#how_it_works">HOW IT WORKS</a>
97
<h1><a name="options">OPTIONS</a></h1>
99
<dt><strong><a name="l_address" class="item"><strong>-l</strong> <em>address</em></a></strong></dt>
102
<p>Tells the daemon to bind to <em>address</em> and accept incoming connections on that
103
socket. If <em>address</em> begins with <code>unix:</code>, everything following that prefix is
104
interpreted as the path to a UNIX domain socket. Otherwise the address or node
105
name are resolved using <code>getaddrinfo()</code>.</p>
106
<p>For network sockets, a port may be specified by using the form
107
<code>[address]:port</code>. If the address is an IPv4 address or a fully
108
qualified domain name (i. e. the address contains at least one dot
109
(<code>.</code>)), the square brackets can be omitted, resulting in the (simpler)
110
<code>address:port</code> pattern. The default port is <strong>42217/udp</strong>. If you
111
specify a network socket, it is mandatory to read the
112
<a href="#security_considerations">SECURITY CONSIDERATIONS</a> section.</p>
113
<p>The following formats are accepted. Please note that the address of the UNIX
114
domain socket <strong>must</strong> start with a slash in the second case!</p>
116
unix:</path/to/unix.sock>
117
/<path/to/unix.sock>
118
<hostname-or-ip>
119
[<hostname-or-ip>]:<port>
120
<hostname-or-ipv4>:<port></pre>
121
<p>If the <strong>-l</strong> option is not specified the default address,
122
<code>unix:/tmp/rrdcached.sock</code>, will be used.</p>
124
<dt><strong><a name="s_group_name_gid" class="item"><strong>-s</strong> <em>group_name</em>|<em>gid</em></a></strong></dt>
127
<p>Set the group permissions of a UNIX domain socket. The option accepts either
128
a numeric group id or group name. That group will then have both read and write
129
permissions (the socket will have file permissions 0750) for the socket and,
130
therefore, is able to send commands to the daemon. This
131
may be useful in cases where you cannot easily run all RRD processes with the same
132
user privileges (e.g. graph generating CGI scripts that typically run in the
133
permission context of the web server).</p>
134
<p>This option affects the <em>following</em> UNIX socket addresses (the following
135
<strong>-l</strong> options), i.e., you may specify different settings for different
137
<p>The default is not to change ownership or permissions of the socket and, thus,
138
use the system default.</p>
140
<dt><strong><a name="m_mode" class="item"><strong>-m</strong> <em>mode</em></a></strong></dt>
143
<p>Set the file permissions of a UNIX domain socket. The option accepts an octal
144
number representing the bit pattern for the mode (see <em>chmod(1)</em> for
146
<p>Please note that not all systems honor this setting. On Linux, read/write
147
permissions are required to connect to a UNIX socket. However, many
148
BSD-derived systems ignore permissions for UNIX sockets. See <em>unix(7)</em> for
150
<p>This option affects the <em>following</em> UNIX socket addresses (the following
151
<strong>-l</strong> options), i.e., you may specify different settings for different
153
<p>The default is not to change ownership or permissions of the socket and, thus,
154
use the system default.</p>
156
<dt><strong><a name="p_command_command" class="item"><strong>-P</strong> <em>command</em>[,<em>command</em>[,...]]</a></strong></dt>
159
<p>Specifies the commands accepted via a network socket. This allows
160
administrators of <em>RRDCacheD</em> to control the actions accepted from various
162
<p>The arguments given to the <strong>-P</strong> option is a comma separated list of commands.
163
For example, to allow the <code>FLUSH</code> and <code>PENDING</code> commands one could specify:</p>
165
rrdcached -P FLUSH,PENDING $MORE_ARGUMENTS</pre>
166
<p>The <strong>-P</strong> option affects the <em>following</em> socket addresses (the following <strong>-l</strong>
167
options). In the following example, only the IPv4 network socket (address
168
<code>10.0.0.1</code>) will be restricted to the <code>FLUSH</code> and <code>PENDING</code> commands:</p>
170
rrdcached -l unix:/some/path -P FLUSH,PENDING -l 10.0.0.1</pre>
171
<p>A complete list of available commands can be found in the section
172
<a href="#valid_commands">Valid Commands</a> below. There are two minor special exceptions:</p>
175
<p>The <code>HELP</code> and <a href="#quit"><code>QUIT</code></a> commands are always allowed.</p>
178
<p>If the <a href="#batch"><code>BATCH</code></a> command is accepted, the <strong>.</strong> command will automatically
179
be accepted, too.</p>
182
<p>Please also read <a href="#security_considerations">SECURITY CONSIDERATIONS</a> below.</p>
184
<dt><strong><a name="w_timeout" class="item"><strong>-w</strong> <em>timeout</em></a></strong></dt>
187
<p>Data is written to disk every <em>timeout</em> seconds. If this option is not
188
specified the default interval of 300 seconds will be used.</p>
190
<dt><strong><a name="z_delay" class="item"><strong>-z</strong> <em>delay</em></a></strong></dt>
193
<p>If specified, rrdcached will delay writing of each RRD for a random number
194
of seconds in the range [0,<em>delay</em>). This will avoid too many
195
writes being queued simultaneously. This value should be no greater than
196
the value specified in <strong>-w</strong>. By default, there is no delay.</p>
198
<dt><strong><a name="f_timeout" class="item"><strong>-f</strong> <em>timeout</em></a></strong></dt>
201
<p>Every <em>timeout</em> seconds the entire cache is searched for old values which are
202
written to disk. This only concerns files to which updates have stopped, so
203
setting this to a high value, such as 3600 seconds, is acceptable in most
204
cases. This timeout defaults to 3600 seconds.</p>
206
<dt><strong><a name="p_file" class="item"><strong>-p</strong> <em>file</em></a></strong></dt>
209
<p>Sets the name and location of the PID-file. If not specified, the default,
210
<code>$localststedir/run/rrdcached.pid</code> will be used.</p>
212
<dt><strong><a name="t_write_threads" class="item"><strong>-t</strong> <em>write_threads</em></a></strong></dt>
215
<p>Specifies the number of threads used for writing RRD files. The default
216
is 4. Increasing this number will allow rrdcached to have more
217
simultaneous I/O requests into the kernel. This may allow the kernel to
218
re-order disk writes, resulting in better disk throughput.</p>
220
<dt><strong><a name="j_dir" class="item"><strong>-j</strong> <em>dir</em></a></strong></dt>
223
<p>Write updates to a journal in <em>dir</em>. In the event of a program or system
224
crash, this will allow the daemon to write any updates that were pending
225
at the time of the crash.</p>
226
<p>On startup, the daemon will check for journal files in this directory. If
227
found, all updates therein will be read into memory before the daemon
228
starts accepting new connections.</p>
229
<p>The journal will be rotated with the same frequency as the flush timer
230
given by <strong>-f</strong>.</p>
231
<p>When journaling is enabled, the daemon will use a fast shutdown procedure.
232
Rather than flushing all files to disk, it will make sure the journal is
233
properly written and exit immediately. Although the RRD data files are
234
not fully up-to-date, no information is lost; all pending updates will be
235
replayed from the journal next time the daemon starts up.</p>
236
<p>To disable fast shutdown, use the <strong>-F</strong> option.</p>
238
<dt><strong><a name="f" class="item"><strong>-F</strong></a></strong></dt>
241
<p>ALWAYS flush all updates to the RRD data files when the daemon is shut
242
down, regardless of journal setting.</p>
244
<dt><strong><a name="g" class="item"><strong>-g</strong></a></strong></dt>
247
<p>Run in the foreground. The daemon will not <code>fork()</code>.</p>
249
<dt><strong><a name="b_dir" class="item"><strong>-b</strong> <em>dir</em></a></strong></dt>
252
<p>The daemon will change into a specific directory at startup. All files passed
253
to the daemon, that are specified by a <strong>relative</strong> path, will be interpreted
254
to be relative to this directory. If not given the default, <code>/tmp</code>, will be
257
+------------------------+------------------------+
258
! Command line ! File updated !
259
+------------------------+------------------------+
260
! foo.rrd ! /tmp/foo.rrd !
261
! foo/bar.rrd ! /tmp/foo/bar.rrd !
262
! /var/lib/rrd/foo.rrd ! /var/lib/rrd/foo.rrd !
263
+------------------------+------------------------+
264
Paths given on the command line and paths actually
265
updated by the daemon, assuming the base directory
266
"/tmp".</pre>
267
<p><strong>WARNING:</strong> The paths up to and including the base directory <strong>MUST NOT BE</strong>
268
symbolic links. In other words, if the base directory is
271
-b /base/dir/somewhere</pre>
272
<p>... then <strong>NONE</strong> of the following should be symbolic links:</p>
276
/base/dir/somewhere</pre>
278
<dt><strong><a name="b" class="item"><strong>-B</strong></a></strong></dt>
281
<p>Only permit writes into the base directory specified in <strong>-b</strong> (and any
282
sub-directories). This does <strong>NOT</strong> detect symbolic links. Paths
283
containing <code>../</code> will also be blocked.</p>
289
<h1><a name="affected_rrdtool_commands">AFFECTED RRDTOOL COMMANDS</a></h1>
290
<p>The following commands may be made aware of the <strong>rrdcached</strong> using the command
291
line argument <strong>--daemon</strong> or the environment variable <strong>RRDCACHED_ADDRESS</strong>:</p>
324
<p>The <strong>update</strong> command can send values to the daemon instead of writing them to
325
the disk itself. All other commands can send a <strong>FLUSH</strong> command (see below) to
326
the daemon before accessing the files, so they work with up-to-date data even
327
if the cache timeout is large.</p>
331
<h1><a name="error_reporting">ERROR REPORTING</a></h1>
332
<p>The daemon reports errors in one of two ways: During startup, error messages
333
are printed to <code>STDERR</code>. One of the steps when starting up is to fork to the
334
background and closing <code>STDERR</code> - after this writing directly to the user is
335
no longer possible. Once this has happened, the daemon will send log messages
336
to the system logging daemon using <code>syslog(3)</code>. The facility used is
337
<code>LOG_DAEMON</code>.</p>
341
<h1><a name="how_it_works">HOW IT WORKS</a></h1>
342
<p>When receiving an update, <strong>rrdcached</strong> does not write to disk but looks for an
343
entry for that file in its internal tree. If not found, an entry is created
344
including the current time (called "First" in the diagram below). This time is
345
<strong>not</strong> the time specified on the command line but the time the operating system
346
considers to be "now". The value and time of the value (called "Time" in the
347
diagram below) are appended to the tree node.</p>
348
<p>When appending a value to a tree node, it is checked whether it's time to write
349
the values to disk. Values are written to disk if
350
<code>now() - First >= timeout</code>, where <code>timeout</code> is the timeout specified
351
using the <strong>-w</strong> option, see <a href="#options">OPTIONS</a>. If the values are "old enough" they
352
will be enqueued in the "update queue", i. e. they will be appended to
353
the linked list shown below. Because the tree nodes and the elements of the
354
linked list are the same data structures in memory, any update to a file that
355
has already been enqueued will be written with the next write to the RRD file,
357
<p>A separate "update thread" constantly dequeues the first element in the update
358
queue and writes all its values to the appropriate file. So as long as the
359
update queue is not empty files are written at the highest possible rate.</p>
360
<p>Since the timeout of files is checked only when new values are added to the
361
file, "dead" files, i. e. files that are not updated anymore, would never
362
be written to disk. Therefore, every now and then, controlled by the <strong>-f</strong>
363
option, the entire tree is walked and all "old" values are enqueued. Since this
364
only affects "dead" files and walking the tree is relatively expensive, you
365
should set the "flush interval" to a reasonably high value. The default is
366
3600 seconds (one hour).</p>
367
<p>The downside of caching values is that they won't show up in graphs generated
368
from the RRD files. To get around this, the daemon provides the "flush
369
command" to flush specific files. This means that the file is inserted at the
370
<strong>head</strong> of the update queue or moved there if it is already enqueued. The flush
371
command will return only after the file's pending updates have been written
374
+------+ +------+ +------+
375
! head ! ! root ! ! tail !
376
+---+--+ +---+--+ +---+--+
380
! /\/\ \ `----------------- ... --------, !
382
+---+----+---+ +------+-----+ +---+----+---+
383
! File: foo ! ! File: bar ! ! File: qux !
384
! First: 101 ! ! First: 119 ! ! First: 180 !
385
! Next:&bar -+--->! Next:&... -+---> ... --->! Next:NULL !
386
| Prev:NULL !<---+-Prev:&foo !<--- ... ----+-Prev: &... !
387
+============+ +============+ +============+
388
! Time: 100 ! ! Time: 120 ! ! Time: 180 !
389
! Value: 10 ! ! Value: 0.1 ! ! Value: 2,2 !
390
+------------+ +------------+ +------------+
391
! Time: 110 ! ! Time: 130 ! ! Time: 190 !
392
! Value: 26 ! ! Value: 0.1 ! ! Value: 7,3 !
393
+------------+ +------------+ +------------+
395
+------------+ +------------+ +------------+
396
! Time: 230 ! ! Time: 250 ! ! Time: 310 !
397
! Value: 42 ! ! Value: 0.2 ! ! Value: 1,2 !
398
+------------+ +------------+ +------------+</pre>
399
<p>The above diagram demonstrates:</p>
402
<p>Files/values are stored in a (balanced) tree.</p>
405
<p>Tree nodes and entries in the update queue are the same data structure.</p>
408
<p>The local time ("First") and the time specified in updates ("Time") may differ.</p>
411
<p>Timed out values are inserted at the "tail".</p>
414
<p>Explicitly flushed values are inserted at the "head".</p>
417
<p>ASCII art rocks.</p>
423
<h1><a name="security_considerations">SECURITY CONSIDERATIONS</a></h1>
426
<h2><a name="authentication">Authentication</a></h2>
427
<p>There is no authentication.</p>
428
<p>The client/server protocol does not yet have any authentication mechanism. It
429
is likely that authentication and encryption will be added in a future version,
430
but for the time being it is the administrator's responsibility to secure the
431
traffic from/to the daemon!</p>
432
<p>It is highly recommended to install a packet filter or similar mechanism to
433
prevent unauthorized connections. Unless you have a dedicated VLAN or VPN for
434
this, using network sockets is probably a bad idea!</p>
437
<h2><a name="authorization">Authorization</a></h2>
438
<p>There is minimal per-socket authorization.</p>
439
<p>Authorization is currently done on a per-socket basis. That means each socket
440
has a list of commands it will accept and it will accept. It will accept only
441
those commands explicitly listed but it will (currently) accept these commands
442
from anyone reaching the socket.</p>
443
<p>If the networking sockets are to be used, it is necessary to restrict the
444
accepted commands to those needed by external clients. If, for example,
445
external clients want to draw graphs of the cached data, they should only be
446
allowed to use the <code>FLUSH</code> command.</p>
449
<h2><a name="encryption">Encryption</a></h2>
450
<p>There is no encryption.</p>
451
<p>Again, this may be added in the future, but for the time being it is your job
452
to keep your private data private. Install a VPN or an encrypted tunnel if you
453
statistics are confidential!</p>
456
<h2><a name="sanity_checking">Sanity checking</a></h2>
457
<p>There is no sanity checking.</p>
458
<p>The daemon will blindly write to any file it gets told, so you really should
459
create a separate user just for this daemon. Also it does not do any sanity
460
checks, so if it gets told to write values for a time far in the future, your
461
files will be messed up good!</p>
464
<h2><a name="conclusion">Conclusion</a></h2>
467
<p>Security is the job of the administrator.</p>
470
<p>We recommend to allow write access via UNIX domain sockets only.</p>
473
<p>You have been warned.</p>
479
<h1><a name="protocol">PROTOCOL</a></h1>
480
<p>The daemon communicates with clients using a line based ASCII protocol which is
481
easy to read and easy to type. This makes it easy for scripts to implement the
482
protocol and possible for users to use telnet to connect to the daemon
483
and test stuff "by hand".</p>
484
<p>The protocol is line based, this means that each record consists of one or more
485
lines. A line is terminated by the line feed character <code>0x0A</code>, commonly
486
written as <code>\n</code>. In the examples below, this character will be written as
487
<code><LF></code> ("line feed").</p>
488
<p>After the connection has been established, the client is expected to send a
489
"command". A command consists of the command keyword, possibly some arguments,
490
and a terminating newline character. For a list of commands, see
491
<a href="#valid_commands">Valid Commands</a> below.</p>
494
FLUSH /tmp/foo.rrd<LF></pre>
495
<p>The daemon answers with a line consisting of a status code and a short status
496
message, separated by one or more space characters. A negative status code
497
signals an error, a positive status code or zero signal success. If the status
498
code is greater than zero, it indicates the number of lines that follow the
502
0 Success<LF></pre>
504
2 Two lines follow<LF>
505
This is the first line<LF>
506
And this is the second line<LF></pre>
509
<h2><a name="valid_commands">Valid Commands</a></h2>
510
<p>The following commands are understood by the daemon:</p>
512
<dt><strong><a name="flush_filename" class="item"><strong>FLUSH</strong> <em>filename</em></a></strong></dt>
515
<p>Causes the daemon to put <em>filename</em> to the <strong>head</strong> of the update queue
516
(possibly moving it there if the node is already enqueued). The answer will be
517
sent <strong>after</strong> the node has been dequeued.</p>
519
<dt><strong><a name="flushall" class="item"><strong>FLUSHALL</strong></a></strong></dt>
522
<p>Causes the daemon to start flushing ALL pending values to disk. This
523
returns immediately, even though the writes may take a long time.</p>
525
<dt><strong><a name="pending_filename" class="item"><strong>PENDING</strong> <em>filename</em></a></strong></dt>
528
<p>Shows any "pending" updates for a file, in order. The updates shown have
529
not yet been written to the underlying RRD file.</p>
531
<dt><strong><a name="forget_filename" class="item"><strong>FORGET</strong> <em>filename</em></a></strong></dt>
534
<p>Removes <em>filename</em> from the cache. Any pending updates <strong>WILL BE LOST</strong>.</p>
536
<dt><strong><a name="queue" class="item"><strong>QUEUE</strong></a></strong></dt>
539
<p>Shows the files that are on the output queue. Returns zero or more lines
540
in the following format, where <num_vals> is the number of values
541
to be written for the <file>:</p>
543
<num_vals> <file></pre>
545
<dt><strong><a name="help_command" class="item"><strong>HELP</strong> [<em>command</em>]</a></strong></dt>
548
<p>Returns a short usage message. If no command is given, or <em>command</em> is
549
<strong>HELP</strong>, a list of commands supported by the daemon is returned. Otherwise a
550
short description, possibly containing a pointer to a manual page, is returned.
551
Obviously, this is meant for interactive usage and the format in which the
552
commands and usage summaries are returned is not well defined.</p>
554
<dt><strong><a name="stats" class="item"><strong>STATS</strong></a></strong></dt>
557
<p>Returns a list of metrics which can be used to measure the daemons performance
558
and check its status. For a description of the values returned, see
559
<a href="#performance_values">Performance Values</a> below.</p>
560
<p>The format in which the values are returned is similar to many other line based
561
protocols: Each value is printed on a separate line, each consisting of the
562
name of the value, a colon, one or more spaces and the actual value.</p>
574
JournalRotate: 0</pre>
576
<dt><strong><a name="update_filename_values_values" class="item"><strong>UPDATE</strong> <em>filename</em> <em>values</em> [<em>values</em> ...]</a></strong></dt>
579
<p>Adds more data to a filename. This is <strong>the</strong> operation the daemon was designed
580
for, so describing the mechanism again is unnecessary. Read <a href="#how_it_works">HOW IT WORKS</a>
581
above for a detailed explanation.</p>
582
<p>Note that rrdcached only accepts absolute timestamps in the update values.
583
Updates strings like "N:1:2:3" are automatically converted to absolute
584
time by the RRD client library before sending to rrdcached.</p>
586
<dt><strong><a name="wrote_filename" class="item"><strong>WROTE</strong> <em>filename</em></a></strong></dt>
589
<p>This command is written to the journal after a file is successfully
590
written out to disk. It is used during journal replay to determine which
591
updates have already been applied. It is <em>only</em> valid in the journal; it
592
is not accepted from the other command channels.</p>
594
<dt><strong><a name="batch" class="item"><strong>BATCH</strong></a></strong></dt>
597
<p>This command initiates the bulk load of multiple commands. This is
598
designed for installations with extremely high update rates, since it
599
permits more than one command to be issued per <code>read()</code> and <code>write()</code>.</p>
600
<p>All commands are executed just as they would be if given individually,
601
except for output to the user. Messages indicating success are
602
suppressed, and error messages are delayed until the client is finished.</p>
603
<p>Command processing is finished when the client sends a dot (".") on its
604
own line. After the client has finished, the server responds with an
605
error count and the list of error messages (if any). Each error messages
606
indicates the number of the command to which it corresponds, and the error
607
message itself. The first user command after <strong>BATCH</strong> is command number one.</p>
610
server: 0 Go ahead. End with dot '.' on its own line.
611
client: UPDATE x.rrd 1223661439:1:2:3 <--- command #1
612
client: UPDATE y.rrd 1223661440:3:4:5 <--- command #2
616
server: 1 message for command 1
617
server: 12 message for command 12</pre>
619
<dt><strong><a name="quit" class="item"><strong>QUIT</strong></a></strong></dt>
622
<p>Disconnect from rrdcached.</p>
627
<h2><a name="performance_values">Performance Values</a></h2>
628
<p>The following counters are returned by the <strong>STATS</strong> command:</p>
630
<dt><strong><a name="queuelength" class="item"><strong>QueueLength</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
633
<p>Number of nodes currently enqueued in the update queue.</p>
635
<dt><strong><a name="updatesreceived" class="item"><strong>UpdatesReceived</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
638
<p>Number of UPDATE commands received.</p>
640
<dt><strong><a name="flushesreceived" class="item"><strong>FlushesReceived</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
643
<p>Number of FLUSH commands received.</p>
645
<dt><strong><a name="updateswritten" class="item"><strong>UpdatesWritten</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
648
<p>Total number of updates, i. e. calls to <code>rrd_update_r</code>, since the
649
daemon was started.</p>
651
<dt><strong><a name="datasetswritten" class="item"><strong>DataSetsWritten</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
654
<p>Total number of "data sets" written to disk since the daemon was
655
started. A data set is one or more values passed to the <strong>UPDATE</strong>
656
command. For example: <code>1223661439:123:456</code> is one data set with two
657
values. The term "data set" is used to prevent confusion whether
658
individual values or groups of values are counted.</p>
660
<dt><strong><a name="treenodesnumber" class="item"><strong>TreeNodesNumber</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
663
<p>Number of nodes in the cache.</p>
665
<dt><strong><a name="treedepth" class="item"><strong>TreeDepth</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
668
<p>Depth of the tree used for fast key lookup.</p>
670
<dt><strong><a name="journalbytes" class="item"><strong>JournalBytes</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
673
<p>Total number of bytes written to the journal since startup.</p>
675
<dt><strong><a name="journalrotate" class="item"><strong>JournalRotate</strong> <em>(unsigned 64bit integer)</em></a></strong></dt>
678
<p>Number of times the journal has been rotated since startup.</p>
684
<h1><a name="signals">SIGNALS</a></h1>
686
<dt><strong><a name="sigint_and_sigterm" class="item">SIGINT and SIGTERM</a></strong></dt>
689
<p>The daemon exits normally on receipt of either of these signals. Pending
690
updates are handled in accordance with the <strong>-j</strong> and <strong>-F</strong> options.</p>
692
<dt><strong><a name="sigusr1" class="item">SIGUSR1</a></strong></dt>
695
<p>The daemon exits AFTER flushing all updates out to disk. This may take a
698
<dt><strong><a name="sigusr2" class="item">SIGUSR2</a></strong></dt>
701
<p>The daemon exits immediately, without flushing updates out to disk.
702
Pending updates will be replayed from the journal when the daemon starts
703
up again. <strong>WARNING: if journaling (-j) is NOT enabled, any pending
704
updates WILL BE LOST</strong>.</p>
710
<h1><a name="bugs">BUGS</a></h1>
711
<p>No known bugs at the moment.</p>
715
<h1><a name="see_also">SEE ALSO</a></h1>
716
<p><a href="././rrdtool.html">the rrdtool manpage</a>, <a href="././rrdgraph.html">the rrdgraph manpage</a></p>
720
<h1><a name="author">AUTHOR</a></h1>
721
<p>Florian Forster <octo at verplant.org></p>
722
<p>Both <strong>rrdcached</strong> and this manual page have been written by Florian.</p>
726
<h1><a name="contributors">CONTRIBUTORS</a></h1>
727
<p>kevin brintnall <<a href="mailto:kbrint@rufus.net">kbrint@rufus.net</a>></p>