Thanks Nate. Switching to Gauge and then recreating the rrd files seems to have fixed the problem.
Thank you! -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Tuesday, July 15, 2008 12:09 PM To: [email protected] Subject: rrd-users Digest, Vol 19, Issue 13 Send rrd-users mailing list submissions to [email protected] To subscribe or unsubscribe via the World Wide Web, visit https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users or, via email, send a message with subject or body 'help' to [EMAIL PROTECTED] You can reach the person managing the list at [EMAIL PROTECTED] When replying, please edit your Subject line so it is more specific than "Re: Contents of rrd-users digest..." Today's Topics: 1. outputting lines and legends in order of weighting (Joseph Mack NA3T) 2. Re: 5 minute data -- RRD (Simon Hobson) 3. Re: outputting lines and legends in order of weighting (Alex van den Bogaerdt) 4. Re: outputting lines and legends in order of weighting (Joseph Mack NA3T) 5. Help with RRDTOL for begginer (Ghaalib Garrin johnson) 6. Newbie question about last_ds and value (Horsky, Laura) 7. Re: Newbie question about last_ds and value (nate) ---------------------------------------------------------------------- Message: 1 Date: Thu, 10 Jul 2008 18:04:10 -0700 (PDT) From: Joseph Mack NA3T <[EMAIL PROTECTED]> Subject: [rrd-users] outputting lines and legends in order of weighting To: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed I'm plotting (among other things), the time offset between the local clock and several ntpd peers. For each peer I plot the jitter as a band of light color (say light red) with the actual time offset as a line (say dark red) in the middle of the band. Since some of the ntp peers are of different stratums (strata?), I plot the data with the least accuracy first, allowing the most accurate data (and jitter bands) to overwrite the least accurate data. All fine so far. However when I plot the boxes in the legend with the colors and value of offset for each ntp peer with the lowest priority lines being written first, the lowest priority peer is written out in the legend first. Is there a way to write out the highest priority legend (and its color) first? I'm happy to do some fakery, eg make two passes at writing out the data (eg one for the lines in one order and another for the legend in the other order), but I can't figure how to do it. Any suggestions? Thanks Joe -- Joseph Mack NA3T EME(B,D), FM05lw North Carolina jmack (at) wm7d (dot) net - azimuthal equidistant map generator at http://www.wm7d.net/azproj.shtml Homepage http://www.austintek.com/ It's GNU/Linux! ------------------------------ Message: 2 Date: Fri, 11 Jul 2008 07:47:38 +0100 From: Simon Hobson <[EMAIL PROTECTED]> Subject: Re: [rrd-users] 5 minute data -- RRD To: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset="us-ascii" ; format="flowed" James Wade wrote: >I am using Hobbit and the RRD database for >monitoring. If I'm collecting 5 minute data, when does the >database begin rolling up the data (averaging it). > >Does it keep a months worth of 5 minute data >and then does a daily average after a month? >Is there a process that compacts the database? >Just wondering how the process works. First off, try Alex's excellent tutorials at http://www.vandenbogaerdt.nl/rrdtool/ But more specifically, each time you do an update, the program checks to see if any time boundaries have been passed. Eg, if you updated every minute for an rrd that has a 5 minute step size, then 4 time it will just accumulate the data, the fifth time (ie when you pass a 5 minute boundary) it will process that step. The consolidations are handled in a similar manner, so a 2 hour consolidation would be processed each time an update occurs that crosses a 2 hour sample period, and so on. There is no database compaction, the rrd database is unique as far as I know in having a fixed size that is set when it is created. Each stored data set is held in a number of preconfigured storage bins - the bins having been allocated at the time the rrd file was created. Each set is treated as a circular buffer (hence the "round robin" in the name), so data is never moved, just the pointer to where the beginning/end is located gets altered along with the contents of that location. ------------------------------ Message: 3 Date: Mon, 14 Jul 2008 01:22:26 +0200 From: Alex van den Bogaerdt <[EMAIL PROTECTED]> Subject: Re: [rrd-users] outputting lines and legends in order of weighting To: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset=us-ascii On Thu, Jul 10, 2008 at 06:04:10PM -0700, Joseph Mack NA3T wrote: > However when I plot the boxes in the legend with the colors > and value of offset for each ntp peer with the lowest > priority lines being written first, the lowest priority peer > is written out in the legend first. Is there a way to write > out the highest priority legend (and its color) first? > > I'm happy to do some fakery, eg make two passes at writing > out the data (eg one for the lines in one order and another > for the legend in the other order), but I can't figure how > to do it. First draw the lines without a legend. Then write a legend without drawing a line. To do this, you need to draw something blank. And this in turn means you first need to generate something blank. If you have: DEF:ds0=test.rrd:ds0:AVERAGE \ DEF:ds1=test.rrd:ds1:AVERAGE \ LINE:ds0#FF0000:ds0 \ LINE:ds1#0000FF:ds1 \ then you change above two LINE in below CDEF and four LINE: CDEF:not=ds0,POP,UNKN \ LINE:ds0#FF0000 \ LINE:ds1#0000FF \ LINE:not#0000FF:ds1 \ LINE:not#FF0000:ds0 \ HTH -- Alex van den Bogaerdt http://www.vandenbogaerdt.nl/rrdtool/ ------------------------------ Message: 4 Date: Sun, 13 Jul 2008 16:36:13 -0700 (PDT) From: Joseph Mack NA3T <[EMAIL PROTECTED]> Subject: Re: [rrd-users] outputting lines and legends in order of weighting To: Alex van den Bogaerdt <[EMAIL PROTECTED]> Cc: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed On Mon, 14 Jul 2008, Alex van den Bogaerdt wrote: > On Thu, Jul 10, 2008 at 06:04:10PM -0700, Joseph Mack NA3T wrote: > >> However when I plot the boxes in the legend with the colors >> and value of offset for each ntp peer with the lowest >> priority lines being written first, the lowest priority peer >> is written out in the legend first. Is there a way to write >> out the highest priority legend (and its color) first? >> >> I'm happy to do some fakery, eg make two passes at writing >> out the data (eg one for the lines in one order and another >> for the legend in the other order), but I can't figure how >> to do it. > > First draw the lines without a legend. > Then write a legend without drawing a line. To do this, you need > to draw something blank. And this in turn means you first need to > generate something blank. thanks I've done a trial run and haven't got data which differentiates my new way from the old way yet. What I've done is draw the error/jitter bands in reverse order, with no legend, then drawn the data lines, in forward order with legend. The lines are narrow enough that I don't really care if the last (lower precision) one overwrites the first one. > > If you have: > > DEF:ds0=test.rrd:ds0:AVERAGE \ > DEF:ds1=test.rrd:ds1:AVERAGE \ > LINE:ds0#FF0000:ds0 \ > LINE:ds1#0000FF:ds1 \ > > then you change above two LINE in below CDEF and four LINE: > > CDEF:not=ds0,POP,UNKN \ > LINE:ds0#FF0000 \ > LINE:ds1#0000FF \ > LINE:not#0000FF:ds1 \ > LINE:not#FF0000:ds0 \ > > HTH Thanks Am keeping my eye on the data for some obvious lower priority jitter bands obliterating the higher priority bands Joe -- Joseph Mack NA3T EME(B,D), FM05lw North Carolina jmack (at) wm7d (dot) net - azimuthal equidistant map generator at http://www.wm7d.net/azproj.shtml Homepage http://www.austintek.com/ It's GNU/Linux! ------------------------------ Message: 5 Date: Mon, 14 Jul 2008 15:03:12 +0200 From: "Ghaalib Garrin johnson" <[EMAIL PROTECTED]> Subject: [rrd-users] Help with RRDTOL for begginer To: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset="iso-8859-1" Hi there I'm still very much a beginner when it comes to rrdtool or mrtg and was wondering if there is anyone out there who has installed rrd on a windows platform and how did they do it. We are currently testing mrtg and we do have stats in the database which we would not like to lose and I want to if it would be possible to move the current mrtg database over to rrdtool on windows 2000 platform. I would really appreciate any help or suggestions regarding this as I would really like to get this software working correctly without losing any information. I'm also having trouble understanding which other software should be run when I have rrdtool installed as well as how I go about installing rrdtool on windows. Thanks and best wishes Ghaalib Johnson -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20080714/366454f 5/attachment-0001.html ------------------------------ Message: 6 Date: Tue, 15 Jul 2008 11:22:47 -0400 From: "Horsky, Laura" <[EMAIL PROTECTED]> Subject: [rrd-users] Newbie question about last_ds and value To: <[email protected]> Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain; charset="us-ascii" I am fairly new to troubleshooting rrd files, so please be patient with me. I am trying to graph the memory usage of a device. I can successfully capture the oid value, but I think it is in the wrong place in the rrd file. What I am seeing is: <ds> <name> memory </name> <type> COUNTER </type> <minimal_heartbeat> 600 </minimal_heartbeat> <min> 0.0000000000e+00 </min> <max> 1.0000000000e+08 </max> <!-- PDP Status --> <last_ds> 260580 </last_ds> <value> 2.1045329206e+09 </value> <unknown_sec> 0 </unknown_sec> </ds> The < last_ds> of 260580 is actual data from the oid, which I thought would also be the "value". What am I doing wrong here? Where is the <value> coming from? This rrd file is being generated in Cacti if that helps and within Cacti I have the following values set: Data source = memory Minimum value = 0 Maximum value = 100000000 (this is artificially high right now while I figure out the value problem) Data Source Type = Counter Heartbeat = 600 Any help would be greatly appreciated. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20080715/291177d 1/attachment-0001.html ------------------------------ Message: 7 Date: Tue, 15 Jul 2008 09:08:47 -0700 (PDT) From: "nate" <[EMAIL PROTECTED]> Subject: Re: [rrd-users] Newbie question about last_ds and value To: [email protected] Message-ID: <[EMAIL PROTECTED]> Content-Type: text/plain;charset=iso-8859-1 Horsky, Laura wrote: > I am fairly new to troubleshooting rrd files, so please be patient with > me. > > > > I am trying to graph the memory usage of a device. I can successfully > capture the oid value, but I think it is in the wrong place in the rrd > file. [..] > <type> COUNTER </type> This should probably be a gauge, of all the different types of hosts I am monitoring memory usage it's always a gauge. nate ------------------------------ _______________________________________________ rrd-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users End of rrd-users Digest, Vol 19, Issue 13 ***************************************** _______________________________________________ rrd-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
