1 RRDTUTORIAL(1) rrdtool RRDTUTORIAL(1)
6 rrdtutorial - Alex van den Bogaerdt's RRDtool tutorial
9 RRDtool is written by Tobias Oetiker <tobi@oetiker.ch> with contribu-
10 tions from many people all around the world. This document is written
11 by Alex van den Bogaerdt <alex@vandenbogaerdt.nl> to help you under-
12 stand what RRDtool is and what it can do for you.
14 The documentation provided with RRDtool can be too technical for some
15 people. This tutorial is here to help you understand the basics of RRD-
16 tool. It should prepare you to read the documentation yourself. It
17 also explains the general things about statistics with a focus on net-
18 working.
23 Please don't skip ahead in this document! The first part of this docu-
24 ment explains the basics and may be boring. But if you don't under-
25 stand the basics, the examples will not be as meaningful to you.
27 Sometimes things change. This example used to provide numbers like
28 "0.04" in stead of "4.00000e-02". Those are really the same numbers,
29 just written down differently. Don't be alarmed if a future version of
30 rrdtool displays a slightly different form of output. The examples in
31 this document are correct for version 1.2.0 of RRDtool.
33 Also, sometimes bugs do occur. They may also influence the outcome of
34 the examples. Example speed4.png was suffering from this (the handling
35 of unknown data in an if-statement was wrong). Normal data will be just
36 fine (a bug in rrdtool wouldn't last long) but special cases like NaN,
37 INF and so on may last a bit longer. Try another version if you can,
38 or just live with it.
40 I fixed the speed4.png example (and added a note). There may be other
41 examples which suffer from the same or a similar bug. Try to fix it
42 yourself, which is a great excercise. But please do not submit your
43 result as a fix to the source of this document. Discuss it on the
44 user's list, or write to me.
48 RRDtool refers to Round Robin Database tool. Round robin is a tech-
49 nique that works with a fixed amount of data, and a pointer to the cur-
50 rent element. Think of a circle with some dots plotted on the edge.
51 These dots are the places where data can be stored. Draw an arrow from
52 the center of the circle to one of the dots; this is the pointer. When
53 the current data is read or written, the pointer moves to the next ele-
54 ment. As we are on a circle there is neither a beginning nor an end,
55 you can go on and on and on. After a while, all the available places
56 will be used and the process automatically reuses old locations. This
57 way, the dataset will not grow in size and therefore requires no main-
58 tenance. RRDtool works with with Round Robin Databases (RRDs). It
59 stores and retrieves data from them.
61 W\bWh\bha\bat\bt d\bda\bat\bta\ba c\bca\ban\bn b\bbe\be p\bpu\but\bt i\bin\bnt\bto\bo a\ban\bn R\bRR\bRD\bD?\b?
63 You name it, it will probably fit as long as it is some sort of time-
64 series data. This means you have to be able to measure some value at
65 several points in time and provide this information to RRDtool. If you
66 can do this, RRDtool will be able to store it. The values must be
67 numerical but don't have to be integers, as is the case with MRTG (the
68 next section will give more details on this more specialized applica-
69 tion).
71 Many examples below talk about SNMP which is an acronym for Simple Net-
72 work Management Protocol. "Simple" refers to the protocol. It does not
73 mean it is simple to manage or monitor a network. After working your
74 way through this document, you should know enough to be able to under-
75 stand what people are talking about. For now, just realize that SNMP
76 can be used to query devices for the values of counters they keep. It
77 is the value from those counters that we want to store in the RRD.
79 W\bWh\bha\bat\bt c\bca\ban\bn I\bI d\bdo\bo w\bwi\bit\bth\bh t\bth\bhi\bis\bs t\bto\boo\bol\bl?\b?
81 RRDtool originated from MRTG (Multi Router Traffic Grapher). MRTG
82 started as a tiny little script for graphing the use of a university's
83 connection to the Internet. MRTG was later (ab-)used as a tool for
84 graphing other data sources including temperature, speed, voltage, num-
85 ber of printouts and the like.
87 Most likely you will start to use RRDtool to store and process data
88 collected via SNMP. The data will most likely be bytes (or bits) trans-
89 fered from and to a network or a computer. But it can also be used to
90 display tidal waves, solar radiation, power consumption, number of vis-
91 itors at an exhibition, noise levels near an airport, temperature on
92 your favorite holiday location, temperature in the fridge and whatever
93 you imagination can come up with.
95 You only need a sensor to measure the data and be able to feed the num-
96 bers into RRDtool. RRDtool then lets you create a database, store data
97 in it, retrieve that data and create graphs in PNG format for display
98 on a web browser. Those PNG images are dependent on the data you col-
99 lected and could be, for instance, an overview of the average network
100 usage, or the peaks that occurred.
102 W\bWh\bha\bat\bt i\bif\bf I\bI s\bst\bti\bil\bll\bl h\bha\bav\bve\be p\bpr\bro\bob\bbl\ble\bem\bms\bs a\baf\bft\bte\ber\br r\bre\bea\bad\bdi\bin\bng\bg t\bth\bhi\bis\bs d\bdo\boc\bcu\bum\bme\ben\bnt\bt?\b?
104 First of all: read it again! You may have missed something. If you are
105 unable to compile the sources and you have a fairly common OS, it will
106 probably not be the fault of RRDtool. There may be pre-compiled ver-
107 sions around on the Internet. If they come from trusted sources, get
108 one of those.
110 If on the other hand the program works but does not give you the
111 expected results, it will be a problem with configuring it. Review your
112 configuration and compare it with the examples that follow.
114 There is a mailing list and an archive of it. Read the list for a few
115 weeks and search the archive. It is considered rude to just ask a ques-
116 tion without searching the archives: your problem may already have been
117 solved for somebody else! This is true for most, if not all, mailing
118 lists and not only for this particular one. Look in the documentation
119 that came with RRDtool for the location and usage of the list.
121 I suggest you take a moment to subscribe to the mailing list right now
122 by sending an email to <rrd-users-request@lists.oetiker.ch> with a sub-
123 ject of "subscribe". If you ever want to leave this list, just write an
124 email to the same address but now with a subject of "unsubscribe".
128 By giving you some detailed descriptions with detailed examples. I
129 assume that following the instructions in the order presented will give
130 you enough knowledge of RRDtool to experiment for yourself. If it
131 doesn't work the first time, don't give up. Reread the stuff that you
132 did understand, you may have missed something.
134 By following the examples you get some hands-on experience and, even
135 more important, some background information of how it works.
137 You will need to know something about hexadecimal numbers. If you don't
138 then start with reading bin_dec_hex before you continue here.
140 Y\bYo\bou\bur\br f\bfi\bir\brs\bst\bt R\bRo\bou\bun\bnd\bd R\bRo\bob\bbi\bin\bn D\bDa\bat\bta\bab\bba\bas\bse\be
142 In my opinion the best way to learn something is to actually do it.
143 Why not start right now? We will create a database, put some values in
144 it and extract this data again. Your output should be the same as the
145 output that is included in this document.
147 We will start with some easy stuff and compare a car with a router, or
148 compare kilometers (miles if you wish) with bits and bytes. It's all
149 the same: some number over some time.
151 Assume we have a device that transfers bytes to and from the Internet.
152 This device keeps a counter that starts at zero when it is turned on,
153 increasing with every byte that is transfered. This counter will proba-
154 bly have a maximum value. If this value is reached and an extra byte is
155 counted, the counter starts over at zero. This is the same as many
156 counters in the world such as the mileage counter in a car.
158 Most discussions about networking talk about bits per second so lets
159 get used to that right away. Assume a byte is eight bits and start to
160 think in bits not bytes. The counter, however, still counts bytes! In
161 the SNMP world most of the counters are 32 bits. That means they are
162 counting from 0 to 4294967295. We will use these values in the exam-
163 ples. The device, when asked, returns the current value of the
164 counter. We know the time that has passes since we last asked so we now
165 know how many bytes have been transfered ***on average*** per second.
166 This is not very hard to calculate. First in words, then in calcula-
167 tions:
169 1. Take the current counter, subtract the previous value from it.
171 2. Do the same with the current time and the previous time (in sec-
172 onds).
174 3. Divide the outcome of (1) by the outcome of (2), the result is the
175 amount of bytes per second. Multiply by eight to get the number of
176 bits per second (bps).
178 bps = (counter_now - counter_before) / (time_now - time_before) * 8
180 For some people it may help to translate this to an automobile example.
181 Do not try this example, and if you do, don't blame me for the results!
183 People who are not used to think in kilometers per hour can translate
184 most into miles per hour by dividing km by 1.6 (close enough). I will
185 use the following abbreviations:
187 m: meter
188 km: kilometer (= 1000 meters).
189 h: hour
190 s: second
191 km/h: kilometers per hour
192 m/s: meters per second
194 You are driving a car. At 12:05 you read the counter in the dashboard
195 and it tells you that the car has moved 12345 km until that moment. At
196 12:10 you look again, it reads 12357 km. This means you have traveled
197 12 km in five minutes. A scientist would translate that into meters per
198 second and this makes a nice comparison toward the problem of (bytes
199 per five minutes) versus (bits per second).
201 We traveled 12 kilometers which is 12000 meters. We did that in five
202 minutes or 300 seconds. Our speed is 12000m / 300s or 40 m/s.
204 We could also calculate the speed in km/h: 12 times 5 minutes is an
205 hour, so we have to multiply 12 km by 12 to get 144 km/h. For our
206 native English speaking friends: that's 90 mph so don't try this exam-
207 ple at home or where I live :)
209 Remember: these numbers are averages only. There is no way to figure
210 out from the numbers, if you drove at a constant speed. There is an
211 example later on in this tutorial that explains this.
213 I hope you understand that there is no difference in calculating m/s or
214 bps; only the way we collect the data is different. Even the k from
215 kilo is the same as in networking terms k also means 1000.
217 We will now create a database where we can keep all these interesting
218 numbers. The method used to start the program may differ slightly from
219 OS to OS, but I assume you can figure it out if it works different on
220 your's. Make sure you do not overwrite any file on your system when
221 executing the following command and type the whole line as one long
222 line (I had to split it for readability) and skip all of the '\' char-
223 acters.
225 rrdtool create test.rrd \
226 --start 920804400 \
227 DS:speed:COUNTER:600:U:U \
228 RRA:AVERAGE:0.5:1:24 \
229 RRA:AVERAGE:0.5:6:10
231 (So enter: "rrdtool create test.rrd --start 920804400 DS ...")
235 We created the round robin database called test (test.rrd) which starts
236 at noon the day I started writing this document, 7th of March, 1999
237 (this date translates to 920804400 seconds as explained below). Our
238 database holds one data source (DS) named "speed" that represents a
239 counter. This counter is read every five minutes (this is the default
240 therefore you don't have to put "--step=300"). In the same database
241 two round robin archives (RRAs) are kept, one averages the data every
242 time it is read (e.g., there's nothing to average) and keeps 24 samples
243 (24 times 5 minutes is 2 hours). The other averages 6 values (half
244 hour) and contains 10 such averages (e.g. 5 hours).
246 RRDtool works with special time stamps coming from the UNIX world.
247 This time stamp is the number of seconds that passed since January 1st
248 1970 UTC. The time stamp value is translated into local time and it
249 will therefore look different for different time zones.
251 Chances are that you are not in the same part of the world as I am.
252 This means your time zone is different. In all examples where I talk
253 about time, the hours may be wrong for you. This has little effect on
254 the results of the examples, just correct the hours while reading. As
255 an example: where I will see "12:05" the UK folks will see "11:05".
257 We now have to fill our database with some numbers. We'll pretend to
258 have read the following numbers:
260 12:05 12345 km
261 12:10 12357 km
262 12:15 12363 km
263 12:20 12363 km
264 12:25 12363 km
265 12:30 12373 km
266 12:35 12383 km
267 12:40 12393 km
268 12:45 12399 km
269 12:50 12405 km
270 12:55 12411 km
271 13:00 12415 km
272 13:05 12420 km
273 13:10 12422 km
274 13:15 12423 km
276 We fill the database as follows:
278 rrdtool update test.rrd 920804700:12345 920805000:12357 920805300:12363
279 rrdtool update test.rrd 920805600:12363 920805900:12363 920806200:12373
280 rrdtool update test.rrd 920806500:12383 920806800:12393 920807100:12399
281 rrdtool update test.rrd 920807400:12405 920807700:12411 920808000:12415
282 rrdtool update test.rrd 920808300:12420 920808600:12422 920808900:12423
284 This reads: update our test database with the following numbers
286 time 920804700, value 12345
287 time 920805000, value 12357
289 etcetera.
291 As you can see, it is possible to feed more than one value into the
292 database in one command. I had to stop at three for readability but the
293 real maximum per line is OS dependent.
295 We can now retrieve the data from our database using "rrdtool fetch":
297 rrdtool fetch test.rrd AVERAGE --start 920804400 --end 920809200
299 It should return the following output:
301 speed
303 920804700: nan
304 920805000: 4.0000000000e-02
305 920805300: 2.0000000000e-02
306 920805600: 0.0000000000e+00
307 920805900: 0.0000000000e+00
308 920806200: 3.3333333333e-02
309 920806500: 3.3333333333e-02
310 920806800: 3.3333333333e-02
311 920807100: 2.0000000000e-02
312 920807400: 2.0000000000e-02
313 920807700: 2.0000000000e-02
314 920808000: 1.3333333333e-02
315 920808300: 1.6666666667e-02
316 920808600: 6.6666666667e-03
317 920808900: 3.3333333333e-03
318 920809200: nan
320 If it doesn't, something may be wrong. Perhaps your OS will print
321 "NaN" in a different form. "NaN" stands for "Not A Number". If your OS
322 writes "U" or "UNKN" or something similar that's okay. If something
323 else is wrong, it will probably be due to an error you made (assuming
324 that my tutorial is correct of course :-). In that case: delete the
325 database and try again.
327 The meaning of the above output will become clear below.
329 T\bTi\bim\bme\be t\bto\bo c\bcr\bre\bea\bat\bte\be s\bso\bom\bme\be g\bgr\bra\bap\bph\bhi\bic\bcs\bs
331 Try the following command:
333 rrdtool graph speed.png \
334 --start 920804400 --end 920808000 \
335 DEF:myspeed=test.rrd:speed:AVERAGE \
336 LINE2:myspeed#FF0000
338 This will create speed.png which starts at 12:00 and ends at 13:00.
339 There is a definition of a variable called myspeed, using the data from
340 RRA "speed" out of database "test.rrd". The line drawn is 2 pixels high
341 and represents the variable myspeed. The color is red (specified by its
342 rgb-representation, see below).
344 You'll notice that the start of the graph is not at 12:00 but at 12:05.
345 This is because we have insufficient data to tell the average before
346 that time. This will only happen when you miss some samples, this will
347 not happen a lot, hopefully.
349 If this has worked: congratulations! If not, check what went wrong.
351 The colors are built up from red, green and blue. For each of the com-
352 ponents, you specify how much to use in hexadecimal where 00 means not
353 included and FF means fully included. The "color" white is a mixture
354 of red, green and blue: FFFFFF The "color" black is all colors off:
355 000000
357 red #FF0000
358 green #00FF00
359 blue #0000FF
360 magenta #FF00FF (mixed red with blue)
361 gray #555555 (one third of all components)
363 Additionally you can (with a recent RRDtool) add an alpha channel
364 (transparency). The default will be "FF" which means non-transparent.
366 The PNG you just created can be displayed using your favorite image
367 viewer. Web browsers will display the PNG via the URL
368 "file:///the/path/to/speed.png"
372 When looking at the image, you notice that the horizontal axis is
373 labeled 12:10, 12:20, 12:30, 12:40 and 12:50. Sometimes a label doesn't
374 fit (12:00 and 13:00 would be likely candidates) so they are skipped.
376 The vertical axis displays the range we entered. We provided kilometers
377 and when divided by 300 seconds, we get very small numbers. To be
378 exact, the first value was 12 (12357-12345) and divided by 300 this
379 makes 0.04, which is displayed by RRDtool as "40 m" meaning "40/1000".
380 The "m" (milli) has nothing to do with meters (also m), kilometers or
381 millimeters! RRDtool doesn't know about the physical units of our data,
382 it just works with dimensionless numbers.
384 If we had measured our distances in meters, this would have been
385 (12357000-12345000)/300 = 12000/300 = 40.
387 As most people have a better feel for numbers in this range, we'll cor-
388 rect that. We could recreate our database and store the correct data,
389 but there is a better way: we do some calculations while creating the
390 png file!
392 rrdtool graph speed2.png \
393 --start 920804400 --end 920808000 \
394 --vertical-label m/s \
395 DEF:myspeed=test.rrd:speed:AVERAGE \
396 CDEF:realspeed=myspeed,1000,\* \
397 LINE2:realspeed#FF0000
399 Note: I need to escape the multiplication operator * with a backslash.
400 If I don't, the operating system may interpret it and use it for file
401 name expansion. You could also place the line within quotation marks
402 like so:
404 "CDEF:realspeed=myspeed,1000,*" \
406 It boils down to: it is RRDtool which should see *, not your shell.
407 And it is your shell interpreting \, not RRDtool. You may need to
408 adjust examples accordingly if you happen to use an operating system or
409 shell which behaves differently.
411 After viewing this PNG, you notice the "m" (milli) has disappeared.
412 This it what the correct result would be. Also, a label has been added
413 to the image. Apart from the things mentioned above, the PNG should
414 look the same.
416 The calculations are specified in the CDEF part above and are in
417 Reverse Polish Notation ("RPN"). What we requested RRDtool to do is:
418 "take the data source myspeed and the number 1000; multiply those".
419 Don't bother with RPN yet, it will be explained later on in more
420 detail. Also, you may want to read my tutorial on CDEFs and Steve
421 Rader's tutorial on RPN. But first finish this tutorial.
423 Hang on! If we can multiply values with 1000, it should also be possi-
424 ble to display kilometers per hour from the same data!
426 To change a value that is measured in meters per second:
428 Calculate meters per hour: value * 3600
429 Calculate kilometers per hour: value / 1000
430 Together this makes: value * (3600/1000) or value * 3.6
432 In our example database we made a mistake and we need to compensate for
433 this by multiplying with 1000. Applying that correction:
435 value * 3.6 * 1000 == value * 3600
437 Now let's create this PNG, and add some more magic ...
439 rrdtool graph speed3.png \
440 --start 920804400 --end 920808000 \
441 --vertical-label km/h \
442 DEF:myspeed=test.rrd:speed:AVERAGE \
443 "CDEF:kmh=myspeed,3600,*" \
444 CDEF:fast=kmh,100,GT,kmh,0,IF \
445 CDEF:good=kmh,100,GT,0,kmh,IF \
446 HRULE:100#0000FF:"Maximum allowed" \
447 AREA:good#00FF00:"Good speed" \
448 AREA:fast#FF0000:"Too fast"
450 Note: here we use another means to escape the * operator by enclosing
451 the whole string in double quotes.
453 This graph looks much better. Speed is shown in km/h and there is even
454 an extra line with the maximum allowed speed (on the road I travel on).
455 I also changed the colors used to display speed and changed it from a
456 line into an area.
458 The calculations are more complex now. For speed measurements within
459 the speed limit they are:
461 Check if kmh is greater than 100 ( kmh,100 ) GT
462 If so, return 0, else kmh ((( kmh,100 ) GT ), 0, kmh) IF
464 For values above the speed limit:
466 Check if kmh is greater than 100 ( kmh,100 ) GT
467 If so, return kmh, else return 0 ((( kmh,100) GT ), kmh, 0) IF
471 I like to believe there are virtually no limits to how RRDtool graph
472 can manipulate data. I will not explain how it works, but look at the
473 following PNG:
475 rrdtool graph speed4.png \
476 --start 920804400 --end 920808000 \
477 --vertical-label km/h \
478 DEF:myspeed=test.rrd:speed:AVERAGE \
479 CDEF:nonans=myspeed,UN,0,myspeed,IF \
480 CDEF:kmh=nonans,3600,* \
481 CDEF:fast=kmh,100,GT,100,0,IF \
482 CDEF:over=kmh,100,GT,kmh,100,-,0,IF \
483 CDEF:good=kmh,100,GT,0,kmh,IF \
484 HRULE:100#0000FF:"Maximum allowed" \
485 AREA:good#00FF00:"Good speed" \
486 AREA:fast#550000:"Too fast" \
487 STACK:over#FF0000:"Over speed"
489 Remember the note in the beginning? I had to remove unknown data from
490 this example. The 'nonans' CDEF is new, and the 6th line (which used to
491 be the 5th line) used to read 'CDEF:kmh=myspeed,3600,*'
493 Let's create a quick and dirty HTML page to view the three PNGs:
495 <HTML><HEAD><TITLE>Speed</TITLE></HEAD><BODY>
496 <IMG src="speed2.png" alt="Speed in meters per second">
497 <BR>
498 <IMG src="speed3.png" alt="Speed in kilometers per hour">
499 <BR>
500 <IMG src="speed4.png" alt="Traveled too fast?">
501 </BODY></HTML>
503 Name the file "speed.html" or similar, and look at it in your web
504 browser.
506 Now, all you have to do is measure the values regularly and update the
507 database. When you want to view the data, recreate the PNGs and make
508 sure to refresh them in your browser. (Note: just clicking reload may
509 not be enough, especially when proxies are involved. Try shift-reload
510 or ctrl-F5).
514 We've already used the "update" command: it took one or more parameters
515 in the form of "<time>:<value>". You'll be glad to know that you can
516 specify the current time by filling in a "N" as the time. Or you could
517 use the "time" function in Perl (the shortest example in this tuto-
518 rial):
520 perl -e 'print time, "\n" '
522 How to run a program on regular intervals is OS specific. But here is
523 an example in pseudo code:
525 - Get the value and put it in variable "$speed"
526 - rrdtool update speed.rrd N:$speed
528 (do not try this with our test database, we'll use it in further
529 examples)
531 This is all. Run the above script every five minutes. When you need to
532 know what the graphs look like, run the examples above. You could put
533 them in a script as well. After running that script, view the page
534 index.html we created above.
538 I can imagine very few people that will be able to get real data from
539 their car every five minutes. All other people will have to settle for
540 some other kind of counter. You could measure the number of pages
541 printed by a printer, for example, the cups of coffee made by the cof-
542 fee machine, a device that counts the electricity used, whatever. Any
543 incrementing counter can be monitored and graphed using the stuff you
544 learned so far. Later on we will also be able to monitor other types of
545 values like temperature.
547 Many people interested in RRDtool will use the counter that keeps track
548 of octets (bytes) transfered by a network device. So let's do just that
549 next. We will start with a description of how to collect data.
551 Some people will make a remark that there are tools which can do this
552 data collection for you. They are right! However, I feel it is impor-
553 tant that you understand they are not necessary. When you have to
554 determine why things went wrong you need to know how they work.
556 One tool used in the example has been talked about very briefly in the
557 beginning of this document, it is called SNMP. It is a way of talking
558 to networked equipment. The tool I use below is called "snmpget" and
559 this is how it works:
561 snmpget device password OID
563 or
565 snmpget -v[version] -c[password] device OID
567 For device you substitute the name, or the IP address, of your device.
568 For password you use the "community read string" as it is called in the
569 SNMP world. For some devices the default of "public" might work, how-
570 ever this can be disabled, altered or protected for privacy and secu-
571 rity reasons. Read the documentation that comes with your device or
572 program.
574 Then there is this parameter, called OID, which means "object identi-
575 fier".
577 When you start to learn about SNMP it looks very confusing. It isn't
578 all that difficult when you look at the Management Information Base
579 ("MIB"). It is an upside-down tree that describes data, with a single
580 node as the root and from there a number of branches. These branches
581 end up in another node, they branch out, etc. All the branches have a
582 name and they form the path that we follow all the way down. The
583 branches that we follow are named: iso, org, dod, internet, mgmt and
584 mib-2. These names can also be written down as numbers and are 1 3 6 1
585 2 1.
587 iso.org.dod.internet.mgmt.mib-2 (1.3.6.1.2.1)
589 There is a lot of confusion about the leading dot that some programs
590 use. There is *no* leading dot in an OID. However, some programs can
591 use the above part of OIDs as a default. To indicate the difference
592 between abbreviated OIDs and full OIDs they need a leading dot when you
593 specify the complete OID. Often those programs will leave out the
594 default portion when returning the data to you. To make things worse,
595 they have several default prefixes ...
597 Ok, lets continue to the start of our OID: we had 1.3.6.1.2.1 From
598 there, we are especially interested in the branch "interfaces" which
599 has number 2 (e.g., 1.3.6.1.2.1.2 or 1.3.6.1.2.1.interfaces).
601 First, we have to get some SNMP program. First look if there is a pre-
602 compiled package available for your OS. This is the preferred way. If
603 not, you will have to get the sources yourself and compile those. The
604 Internet is full of sources, programs etc. Find information using a
605 search engine or whatever you prefer.
607 Assume you got the program. First try to collect some data that is
608 available on most systems. Remember: there is a short name for the part
609 of the tree that interests us most in the world we live in!
611 I will give an example which can be used on Fedora Core 3. If it
612 doesn't work for you, work your way through the manual of snmp and
613 adapt the example to make it work.
615 snmpget -v2c -c public myrouter system.sysDescr.0
617 The device should answer with a description of itself, perhaps an empty
618 one. Until you got a valid answer from a device, perhaps using a dif-
619 ferent "password", or a different device, there is no point in continu-
620 ing.
622 snmpget -v2c -c public myrouter interfaces.ifNumber.0
624 Hopefully you get a number as a result, the number of interfaces. If
625 so, you can carry on and try a different program called "snmpwalk".
627 snmpwalk -v2c -c public myrouter interfaces.ifTable.ifEntry.ifDescr
629 If it returns with a list of interfaces, you're almost there. Here's
630 an example:
631 [user@host /home/alex]$ snmpwalk -v2c -c public cisco 2.2.1.2
633 interfaces.ifTable.ifEntry.ifDescr.1 = "BRI0: B-Channel 1"
634 interfaces.ifTable.ifEntry.ifDescr.2 = "BRI0: B-Channel 2"
635 interfaces.ifTable.ifEntry.ifDescr.3 = "BRI0" Hex: 42 52 49 30
636 interfaces.ifTable.ifEntry.ifDescr.4 = "Ethernet0"
637 interfaces.ifTable.ifEntry.ifDescr.5 = "Loopback0"
639 On this cisco equipment, I would like to monitor the "Ethernet0" inter-
640 face and from the above output I see that it is number four. I try:
642 [user@host /home/alex]$ snmpget -v2c -c public cisco 2.2.1.10.4 2.2.1.16.4
644 interfaces.ifTable.ifEntry.ifInOctets.4 = 2290729126
645 interfaces.ifTable.ifEntry.ifOutOctets.4 = 1256486519
647 So now I have two OIDs to monitor and they are (in full, this time):
649 1.3.6.1.2.1.2.2.1.10
651 and
653 1.3.6.1.2.1.2.2.1.16
655 both with an interface number of 4.
657 Don't get fooled, this wasn't my first try. It took some time for me
658 too to understand what all these numbers mean. It does help a lot when
659 they get translated into descriptive text... At least, when people are
660 talking about MIBs and OIDs you know what it's all about. Do not
661 forget the interface number (0 if it is not interface dependent) and
662 try snmpwalk if you don't get an answer from snmpget.
664 If you understand the above section and get numbers from your device,
665 continue on with this tutorial. If not, then go back and re-read this
666 part.
670 Let the fun begin. First, create a new database. It contains data from
671 two counters, called input and output. The data is put into archives
672 that average it. They take 1, 6, 24 or 288 samples at a time. They
673 also go into archives that keep the maximum numbers. This will be
674 explained later on. The time in-between samples is 300 seconds, a good
675 starting point, which is the same as five minutes.
677 1 sample "averaged" stays 1 period of 5 minutes
678 6 samples averaged become one average on 30 minutes
679 24 samples averaged become one average on 2 hours
680 288 samples averaged become one average on 1 day
682 Lets try to be compatible with MRTG which stores about the following
683 amount of data:
685 600 5-minute samples: 2 days and 2 hours
686 600 30-minute samples: 12.5 days
687 600 2-hour samples: 50 days
688 732 1-day samples: 732 days
690 These ranges are appended, so the total amount of data stored in the
691 database is approximately 797 days. RRDtool stores the data differ-
692 ently, it doesn't start the "weekly" archive where the "daily" archive
693 stopped. For both archives the most recent data will be near "now" and
694 therefore we will need to keep more data than MRTG does!
696 We will need:
698 600 samples of 5 minutes (2 days and 2 hours)
699 700 samples of 30 minutes (2 days and 2 hours, plus 12.5 days)
700 775 samples of 2 hours (above + 50 days)
701 797 samples of 1 day (above + 732 days, rounded up to 797)
703 rrdtool create myrouter.rrd \
704 DS:input:COUNTER:600:U:U \
705 DS:output:COUNTER:600:U:U \
706 RRA:AVERAGE:0.5:1:600 \
707 RRA:AVERAGE:0.5:6:700 \
708 RRA:AVERAGE:0.5:24:775 \
709 RRA:AVERAGE:0.5:288:797 \
710 RRA:MAX:0.5:1:600 \
711 RRA:MAX:0.5:6:700 \
712 RRA:MAX:0.5:24:775 \
713 RRA:MAX:0.5:288:797
715 Next thing to do is to collect data and store it. Here is an example.
716 It is written partially in pseudo code, you will have to find out what
717 to do exactly on your OS to make it work.
719 while not the end of the universe
720 do
721 get result of
722 snmpget router community 2.2.1.10.4
723 into variable $in
724 get result of
725 snmpget router community 2.2.1.16.4
726 into variable $out
728 rrdtool update myrouter.rrd N:$in:$out
730 wait for 5 minutes
731 done
733 Then, after collecting data for a day, try to create an image using:
735 rrdtool graph myrouter-day.png --start -86400 \
736 DEF:inoctets=myrouter.rrd:input:AVERAGE \
737 DEF:outoctets=myrouter.rrd:output:AVERAGE \
738 AREA:inoctets#00FF00:"In traffic" \
739 LINE1:outoctets#0000FF:"Out traffic"
741 This should produce a picture with one day worth of traffic. One day
742 is 24 hours of 60 minutes of 60 seconds: 24*60*60=86400, we start at
743 now minus 86400 seconds. We define (with DEFs) inoctets and outoctets
744 as the average values from the database myrouter.rrd and draw an area
745 for the "in" traffic and a line for the "out" traffic.
747 View the image and keep logging data for a few more days. If you like,
748 you could try the examples from the test database and see if you can
749 get various options and calculations to work.
751 Suggestion: Display in bytes per second and in bits per second. Make
752 the Ethernet graphics go red if they are over four megabits per second.
756 A few paragraphs back I mentioned the possibility of keeping the maxi-
757 mum values instead of the average values. Let's go into this a bit
758 more.
760 Recall all the stuff about the speed of the car. Suppose we drove at
761 144 km/h during 5 minutes and then were stopped by the police for 25
762 minutes. At the end of the lecture we would take our laptop and create
763 and view the image taken from the database. If we look at the second
764 RRA we did create, we would have the average from 6 samples. The sam-
765 ples measured would be 144+0+0+0+0+0=144, divided by 30 minutes, cor-
766 rected for the error by 1000, translated into km/h, with a result of 24
767 km/h. I would still get a ticket but not for speeding anymore :)
769 Obviously, in this case we shouldn't look at the averages. In some
770 cases they are handy. If you want to know how many km you had traveled,
771 the averaged picture would be the right one to look at. On the other
772 hand, for the speed that we traveled at, the maximum numbers seen is
773 much more interesting. Later we will see more types.
775 It is the same for data. If you want to know the amount, look at the
776 averages. If you want to know the rate, look at the maximum. Over
777 time, they will grow apart more and more. In the last database we have
778 created, there are two archives that keep data per day. The archive
779 that keeps averages will show low numbers, the archive that shows max-
780 ima will have higher numbers.
782 For my car this would translate in averages per day of 96/24=4 km/h (as
783 I travel about 94 kilometers on a day) during working days, and maxima
784 of 120 km/h (my top speed that I reach every day).
786 Big difference. Do not look at the second graph to estimate the dis-
787 tances that I travel and do not look at the first graph to estimate my
788 speed. This will work if the samples are close together, as they are in
789 five minutes, but not if you average.
791 On some days, I go for a long ride. If I go across Europe and travel
792 for 12 hours, the first graph will rise to about 60 km/h. The second
793 one will show 180 km/h. This means that I traveled a distance of 60
794 km/h times 24 h = 1440 km. I did this with a higher speed and a maximum
795 around 180 km/h. However, it probably doesn't mean that I traveled for
796 8 hours at a constant speed of 180 km/h!
798 This is a real example: go with the flow through Germany (fast!) and
799 stop a few times for gas and coffee. Drive slowly through Austria and
800 the Netherlands. Be careful in the mountains and villages. If you would
801 look at the graphs created from the five-minute averages you would get
802 a totally different picture. You would see the same values on the aver-
803 age and maximum graphs (provided I measured every 300 seconds). You
804 would be able to see when I stopped, when I was in top gear, when I
805 drove over fast highways etc. The granularity of the data is much
806 higher, so you can see more. However, this takes 12 samples per hour,
807 or 288 values per day, so it would be a lot of data over a longer
808 period of time. Therefore we average it, eventually to one value per
809 day. From this one value, we cannot see much detail, of course.
811 Make sure you understand the last few paragraphs. There is no value in
812 only a line and a few axis, you need to know what they mean and inter-
813 pret the data in an appropriate way. This is true for all data.
815 The biggest mistake you can make is to use the collected data for some-
816 thing that it is not suitable for. You would be better off if you
817 didn't have the graph at all.
819 L\bLe\bet\bt'\b's\bs r\bre\bev\bvi\bie\bew\bw w\bwh\bha\bat\bt y\byo\bou\bu n\bno\bow\bw s\bsh\bho\bou\bul\bld\bd k\bkn\bno\bow\bw
821 You know how to create a database and can put data in it. You can get
822 the numbers out again by creating an image, do math on the data from
823 the database and view the result instead of the raw data. You know
824 about the difference between averages and maximum, and when to use
825 which (or at least you should have an idea).
827 RRDtool can do more than what we have learned up to now. Before you
828 continue with the rest of this doc, I recommend that you reread from
829 the start and try some modifications on the examples. Make sure you
830 fully understand everything. It will be worth the effort and helps you
831 not only with the rest of this tutorial, but also in your day to day
832 monitoring long after you read this introduction.
836 All right, you feel like continuing. Welcome back and get ready for an
837 increased speed in the examples and explanations.
839 You know that in order to view a counter over time, you have to take
840 two numbers and divide the difference of them between the time lapsed.
841 This makes sense for the examples I gave you but there are other possi-
842 bilities. For instance, I'm able to retrieve the temperature from my
843 router in three places namely the inlet, the so called hot-spot and the
844 exhaust. These values are not counters. If I take the difference of
845 the two samples and divide that by 300 seconds I would be asking for
846 the temperature change per second. Hopefully this is zero! If not, the
847 computer room is probably on fire :)
849 So, what can we do? We can tell RRDtool to store the values we measure
850 directly as they are (this is not entirely true but close enough). The
851 graphs we make will look much better, they will show a rather constant
852 value. I know when the router is busy (it works -> it uses more elec-
853 tricity -> it generates more heat -> the temperature rises). I know
854 when the doors are left open (the room is air conditioned) -> the warm
855 air from the rest of the building flows into the computer room -> the
856 inlet temperature rises). Etc. The data type we use when creating the
857 database before was counter, we now have a different data type and thus
858 a different name for it. It is called GAUGE. There are more such data
859 types:
861 - COUNTER we already know this one
862 - GAUGE we just learned this one
863 - DERIVE
864 - ABSOLUTE
866 The two additional types are DERIVE and ABSOLUTE. Absolute can be used
867 like counter with one difference: RRDtool assumes the counter is reset
868 when it's read. That is: its delta is known without calculation by RRD-
869 tool whereas RRDtool needs to calculate it for the counter type. Exam-
870 ple: our first example (12345, 12357, 12363, 12363) would read:
871 unknown, 12, 6, 0. The rest of the calculations stay the same. The
872 other one, derive, is like counter. Unlike counter, it can also
873 decrease so it can have a negative delta. Again, the rest of the calcu-
874 lations stay the same.
876 Let's try them all:
878 rrdtool create all.rrd --start 978300900 \
879 DS:a:COUNTER:600:U:U \
880 DS:b:GAUGE:600:U:U \
881 DS:c:DERIVE:600:U:U \
882 DS:d:ABSOLUTE:600:U:U \
883 RRA:AVERAGE:0.5:1:10
884 rrdtool update all.rrd \
885 978301200:300:1:600:300 \
886 978301500:600:3:1200:600 \
887 978301800:900:5:1800:900 \
888 978302100:1200:3:2400:1200 \
889 978302400:1500:1:2400:1500 \
890 978302700:1800:2:1800:1800 \
891 978303000:2100:4:0:2100 \
892 978303300:2400:6:600:2400 \
893 978303600:2700:4:600:2700 \
894 978303900:3000:2:1200:3000
895 rrdtool graph all1.png -s 978300600 -e 978304200 -h 400 \
896 DEF:linea=all.rrd:a:AVERAGE LINE3:linea#FF0000:"Line A" \
897 DEF:lineb=all.rrd:b:AVERAGE LINE3:lineb#00FF00:"Line B" \
898 DEF:linec=all.rrd:c:AVERAGE LINE3:linec#0000FF:"Line C" \
899 DEF:lined=all.rrd:d:AVERAGE LINE3:lined#000000:"Line D"
901 R\bRR\bRD\bDt\bto\boo\bol\bl u\bun\bnd\bde\ber\br t\bth\bhe\be M\bMi\bic\bcr\bro\bos\bsc\bco\bop\bpe\be
904 · Line A is a COUNTER type, so it should continuously increment and
905 RRDtool must calculate the differences. Also, RRDtool needs to divide
906 the difference by the amount of time lapsed. This should end up as a
907 straight line at 1 (the deltas are 300, the time is 300).
909 · Line B is of type GAUGE. These are "real" values so they should match
910 what we put in: a sort of a wave.
912 · Line C is of type DERIVE. It should be a counter that can decrease.
913 It does so between 2400 and 0, with 1800 in-between.
915 · Line D is of type ABSOLUTE. This is like counter but it works on val-
916 ues without calculating the difference. The numbers are the same and
917 as you can see (hopefully) this has a different result.
919 This translates in the following values, starting at 23:10 and ending
920 at 00:10 the next day (where "u" means unknown/unplotted):
922 - Line A: u u 1 1 1 1 1 1 1 1 1 u
923 - Line B: u 1 3 5 3 1 2 4 6 4 2 u
924 - Line C: u u 2 2 2 0 -2 -6 2 0 2 u
925 - Line D: u 1 2 3 4 5 6 7 8 9 10 u
927 If your PNG shows all this, you know you have entered the data cor-
928 rectly, the RRDtool executable is working properly, your viewer doesn't
929 fool you, and you successfully entered the year 2000 :)
931 You could try the same example four times, each time with only one of
932 the lines.
934 Let's go over the data again:
936 · Line A: 300,600,900 and so on. The counter delta is a constant 300
937 and so is the time delta. A number divided by itself is always 1
938 (except when dividing by zero which is undefined/illegal).
940 Why is it that the first point is unknown? We do know what we put
941 into the database, right? True, But we didn't have a value to calcu-
942 late the delta from, so we don't know where we started. It would be
943 wrong to assume we started at zero so we don't!
945 · Line B: There is nothing to calculate. The numbers are as they are.
947 · Line C: Again, the start-out value is unknown. The same story is
948 holds as for line A. In this case the deltas are not constant, there-
949 fore the line is not either. If we would put the same numbers in the
950 database as we did for line A, we would have gotten the same line.
951 Unlike type counter, this type can decrease and I hope to show you
952 later on why this makes a difference.
954 · Line D: Here the device calculates the deltas. Therefore we DO know
955 the first delta and it is plotted. We had the same input as with line
956 A, but the meaning of this input is different and thus the line is
957 different. In this case the deltas increase each time with 300. The
958 time delta stays at a constant 300 and therefore the division of the
959 two gives increasing values.
963 There are a few more basics to show. Some important options are still
964 to be covered and we haven't look at counter wraps yet. First the
965 counter wrap: In our car we notice that the counter shows 999987. We
966 travel 20 km and the counter should go to 1000007. Unfortunately, there
967 are only six digits on our counter so it really shows 000007. If we
968 would plot that on a type DERIVE, it would mean that the counter was
969 set back 999980 km. It wasn't, and there has to be some protection for
970 this. This protection is only available for type COUNTER which should
971 be used for this kind of counter anyways. How does it work? Type
972 counter should never decrease and therefore RRDtool must assume it
973 wrapped if it does decrease! If the delta is negative, this can be
974 compensated for by adding the maximum value of the counter + 1. For our
975 car this would be:
977 Delta = 7 - 999987 = -999980 (instead of 1000007-999987=20)
979 Real delta = -999980 + 999999 + 1 = 20
981 At the time of writing this document, RRDtool knows of counters that
982 are either 32 bits or 64 bits of size. These counters can handle the
983 following different values:
985 - 32 bits: 0 .. 4294967295
986 - 64 bits: 0 .. 18446744073709551615
988 If these numbers look strange to you, you can view them in their hex-
989 adecimal form:
991 - 32 bits: 0 .. FFFFFFFF
992 - 64 bits: 0 .. FFFFFFFFFFFFFFFF
994 RRDtool handles both counters the same. If an overflow occurs and the
995 delta would be negative, RRDtool first adds the maximum of a small
996 counter + 1 to the delta. If the delta is still negative, it had to be
997 the large counter that wrapped. Add the maximum possible value of the
998 large counter + 1 and subtract the erroneously added small value.
1000 There is a risk in this: suppose the large counter wrapped while adding
1001 a huge delta, it could happen, theoretically, that adding the smaller
1002 value would make the delta positive. In this unlikely case the results
1003 would not be correct. The increase should be nearly as high as the max-
1004 imum counter value for that to happen, so chances are you would have
1005 several other problems as well and this particular problem would not
1006 even be worth thinking about. Even though, I did include an example, so
1007 you can judge for yourself.
1009 The next section gives you some numerical examples for counter-wraps.
1010 Try to do the calculations yourself or just believe me if your calcula-
1011 tor can't handle the numbers :)
1013 Correction numbers:
1015 - 32 bits: (4294967295 + 1) = 4294967296
1016 - 64 bits: (18446744073709551615 + 1)
1017 - correction1 = 18446744069414584320
1019 Before: 4294967200
1020 Increase: 100
1021 Should become: 4294967300
1022 But really is: 4
1023 Delta: -4294967196
1024 Correction1: -4294967196 + 4294967296 = 100
1026 Before: 18446744073709551000
1027 Increase: 800
1028 Should become: 18446744073709551800
1029 But really is: 184
1030 Delta: -18446744073709550816
1031 Correction1: -18446744073709550816
1032 + 4294967296 = -18446744069414583520
1033 Correction2: -18446744069414583520
1034 + 18446744069414584320 = 800
1036 Before: 18446744073709551615 ( maximum value )
1037 Increase: 18446744069414584320 ( absurd increase, minimum for
1038 Should become: 36893488143124135935 this example to work )
1039 But really is: 18446744069414584319
1040 Delta: -4294967296
1041 Correction1: -4294967296 + 4294967296 = 0
1042 (not negative -> no correction2)
1044 Before: 18446744073709551615 ( maximum value )
1045 Increase: 18446744069414584319 ( one less increase )
1046 Should become: 36893488143124135934
1047 But really is: 18446744069414584318
1048 Delta: -4294967297
1049 Correction1: -4294967297 + 4294967296 = -1
1050 Correction2: -1 + 18446744069414584320 = 18446744069414584319
1052 As you can see from the last two examples, you need strange numbers for
1053 RRDtool to fail (provided it's bug free of course), so this should not
1054 happen. However, SNMP or whatever method you choose to collect the
1055 data, might also report wrong numbers occasionally. We can't prevent
1056 all errors, but there are some things we can do. The RRDtool "create"
1057 command takes two special parameters for this. They define the minimum
1058 and maximum allowed values. Until now, we used "U", meaning "unknown".
1059 If you provide values for one or both of them and if RRDtool receives
1060 data points that are outside these limits, it will ignore those values.
1061 For a thermometer in degrees Celsius, the absolute minimum is just
1062 under -273. For my router, I can assume this minimum is much higher so
1063 I would set it to 10, where as the maximum temperature I would set to
1064 80. Any higher and the device would be out of order.
1066 For the speed of my car, I would never expect negative numbers and also
1067 I would not expect a speed higher than 230. Anything else, and there
1068 must have been an error. Remember: the opposite is not true, if the
1069 numbers pass this check, it doesn't mean that they are correct. Always
1070 judge the graph with a healthy dose of suspicion if it seems weird to
1071 you.
1075 One important feature of RRDtool has not been explained yet: it is vir-
1076 tually impossible to collect data and feed it into RRDtool on exact
1077 intervals. RRDtool therefore interpolates the data, so they are stored
1078 on exact intervals. If you do not know what this means or how it works,
1079 then here's the help you seek:
1081 Suppose a counter increases by exactly one for every second. You want
1082 to measure it in 300 seconds intervals. You should retrieve values that
1083 are exactly 300 apart. However, due to various circumstances you are a
1084 few seconds late and the interval is 303. The delta will also be 303 in
1085 that case. Obviously, RRDtool should not put 303 in the database and
1086 make you believe that the counter increased by 303 in 300 seconds.
1087 This is where RRDtool interpolates: it alters the 303 value as if it
1088 would have been stored earlier and it will be 300 in 300 seconds. Next
1089 time you are at exactly the right time. This means that the current
1090 interval is 297 seconds and also the counter increased by 297. Again,
1091 RRDtool interpolates and stores 300 as it should be.
1093 in the RRD in reality
1095 time+000: 0 delta="U" time+000: 0 delta="U"
1096 time+300: 300 delta=300 time+300: 300 delta=300
1097 time+600: 600 delta=300 time+603: 603 delta=303
1098 time+900: 900 delta=300 time+900: 900 delta=297
1100 Let's create two identical databases. I've chosen the time range
1101 920805000 to 920805900 as this goes very well with the example numbers.
1103 rrdtool create seconds1.rrd \
1104 --start 920804700 \
1105 DS:seconds:COUNTER:600:U:U \
1106 RRA:AVERAGE:0.5:1:24
1108 Make a copy
1110 for Unix: cp seconds1.rrd seconds2.rrd
1111 for Dos: copy seconds1.rrd seconds2.rrd
1112 for vms: how would I know :)
1114 Put in some data
1116 rrdtool update seconds1.rrd \
1117 920805000:000 920805300:300 920805600:600 920805900:900
1118 rrdtool update seconds2.rrd \
1119 920805000:000 920805300:300 920805603:603 920805900:900
1121 Create output
1123 rrdtool graph seconds1.png \
1124 --start 920804700 --end 920806200 \
1125 --height 200 \
1126 --upper-limit 1.05 --lower-limit 0.95 --rigid \
1127 DEF:seconds=seconds1.rrd:seconds:AVERAGE \
1128 CDEF:unknown=seconds,UN \
1129 LINE2:seconds#0000FF \
1130 AREA:unknown#FF0000
1131 rrdtool graph seconds2.png \
1132 --start 920804700 --end 920806200 \
1133 --height 200 \
1134 --upper-limit 1.05 --lower-limit 0.95 --rigid \
1135 DEF:seconds=seconds2.rrd:seconds:AVERAGE \
1136 CDEF:unknown=seconds,UN \
1137 LINE2:seconds#0000FF \
1138 AREA:unknown#FF0000
1140 View both images together (add them to your index.html file) and com-
1141 pare. Both graphs should show the same, despite the input being differ-
1142 ent.
1145 It's time now to wrap up this tutorial. We covered all the basics for
1146 you to be able to work with RRDtool and to read the additional documen-
1147 tation available. There is plenty more to discover about RRDtool and
1148 you will find more and more uses for this package. You can easily cre-
1149 ate graphs using just the examples provided and using only RRDtool. You
1150 can also use one of the front ends to RRDtool that are available.
1153 Remember to subscribe to the RRDtool mailing list. Even if you are not
1154 answering to mails that come by, it helps both you and the rest of the
1155 users. A lot of the stuff that I know about MRTG (and therefore about
1156 RRDtool) I've learned while just reading the list without posting to
1157 it. I did not need to ask the basic questions as they are answered in
1158 the FAQ (read it!) and in various mails by other users. With thousands
1159 of users all over the world, there will always be people who ask ques-
1160 tions that you can answer because you read this and other documentation
1161 and they didn't.
1164 The RRDtool manpages
1167 I hope you enjoyed the examples and their descriptions. If you do, help
1168 other people by pointing them to this document when they are asking
1169 basic questions. They will not only get their answers, but at the same
1170 time learn a whole lot more.
1172 Alex van den Bogaerdt <alex@vandenbogaerdt.nl>
1176 1.3.99909060808 2009-04-09 RRDTUTORIAL(1)