<div class="gmail_extra">I had no response on this. What you've reported is Interesting. I run RRDtool v1.2.12. I might be able to upgrade RRDtool on a test server to see if it makes a difference, but not for a few days at least.</div>
<div class="gmail_extra"><br></div><div class="gmail_extra">I'm a little skeptical of the theory that rrdtool could be the difference, but only because I believe Xymon is generating the URL with the "start" and "count" parameters, and doesn't seem to consult rrdtool for any start/count values or similar. Furthermore, there doesn't seem to be a way to query rrdtool for these values. But this is conjecture, as it's been a while since I reviewed the code.</div>
<div class="gmail_extra"><br></div><div class="gmail_extra">J<br><br><div class="gmail_quote">On Sun, Apr 29, 2012 at 4:02 AM, Ralph Mitchell <span dir="ltr"><<a href="mailto:ralphmitchell@gmail.com" target="_blank">ralphmitchell@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Here's a twist - I just tried it at home on a Gentoo system and<br>
xymon-4.3.7 is happily doing all the graphs for 11 filesystems,<br>
distributed as 3 + 3 + 3 + 2. When I mouse-over the 2 disk graph, the<br>
url shows "start=10&count=3", so maybe it's a problem with rrdtool? I<br>
know there are other things in between that url and actually getting<br>
the graph, but if rrdtool is told to graph 3 data sources and is only<br>
given 2, it could just be choking.<br>
<br>
I don't know what version I have in SLES11 at work, but at home it's<br>
rrdtool-1.4.5-r1. Maybe this version handles it more gracefully if<br>
there's a data source missing.<br>
<span class="HOEnZb"><font color="#888888"><br>
Ralph Mitchell<br>
</font></span><div class="HOEnZb"><div class="h5"><br>
<br>
On Thu, Apr 26, 2012 at 7:46 PM, Ralph Mitchell <<a href="mailto:ralphmitchell@gmail.com">ralphmitchell@gmail.com</a>> wrote:<br>
> Did you ever get any response on this broken graph issue? I'm seeing the<br>
> same thing now.<br>
><br>
> I have a system with 11 file systems. Multi-graph shows them in sets of 3,<br>
> except for the last 2, which don't show at all. When I view the page<br>
> source, I can see that the link says first=10&count=3. I figure that must be<br>
> breaking the graph generator because there are only 2 more graphs to show.<br>
><br>
> This is stock 4.3.7, btw. I'll dig into the code later and see if I can see<br>
> why the math miscounts.<br>
><br>
> Ralph Mitchell<br>
><br>
> On Mar 15, 2012 11:14 PM, "Jeremy Laidman" <<a href="mailto:jlaidman@rebel-it.com.au">jlaidman@rebel-it.com.au</a>> wrote:<br>
>><br>
>> Xymonsters<br>
>><br>
>> I'm looking for a way forward on the "broken multiple graphs" problem<br>
>> that I'm experiencing. Excuse the long post; it's part brainstorming,<br>
>> and part context-setting.<br>
>><br>
>> For a while there have been some discussion on the list about how<br>
>> custom graphs can get displayed incorrectly as if they're multi-graph<br>
>> when they're not. (I think I'm using the correct term "multi-graph"<br>
>> here.) Xymon works out that there are enough RRD files to fill more<br>
>> than one graph, and shows several graphs, each with a different<br>
>> "first" parameter. For me, in many cases all of the lines fit on one<br>
>> graph, and so the other graphs show as a broken image icon.<br>
>><br>
>> A work-around seems to be to explicitly show "<!-- linecount=N>" to<br>
>> define how many lines should be expected on a set of graphs. I've<br>
>> done this in my devmon templates to prevent the broken graphs.<br>
>><br>
>> However, this work-around is only available for custom graphs that<br>
>> create their own status webpage using HTML (in which the "<!--<br>
>> linecount=N>" tag can be included) and send them to Xymon using a<br>
>> "status" message.<br>
>><br>
>> I have some collectors that send their data to Xymon using "data"<br>
>> messages, because I don't want to grow my Xymon display too wide, with<br>
>> a bunch of dots that can only ever be green. So my graphs show up in<br>
>> the trends column, which is exactly where I want them. But because I<br>
>> can't add a "linecount" anywhere, I get broken graphs.<br>
>><br>
>> I really don't think I'm doing anything out of the ordinary here, and<br>
>> certainly not deviating from the documentation. I've scoured doco,<br>
>> mailing list, and even code to find out what can be done to work<br>
>> around this problem. Finally, I've come to the conclusion that the<br>
>> Xymon code is simply unable to work out how many graphs to show in all<br>
>> situations, and its heuristics are not working sometimes. This is the<br>
>> only bit of my Xymon installation that is tarnishing its reputation<br>
>> here, meaning that it doesn't compare favourably with other products,<br>
>> so I'd really like to get this fixed somehow. But I'm not sure what<br>
>> to do about it. If I'm simply doing something wrong, then please let<br>
>> me know, and ignore the rest of this post.<br>
>><br>
>> So, what am I asking for? There might be a way to get Xymon to work<br>
>> out the graph linecount more universally, perhaps at the expense of<br>
>> more code complexity, compatibility, CPU cycles or disk I/O (I think<br>
>> the some code comments describe an intent to avoid scanning the<br>
>> directory). But even so, I think what I'd like to have is a way of<br>
>> overriding the heuristics other than the "linecount" hack. But what's<br>
>> the best way to do this? I'm willing to cut code to get this fixed,<br>
>> but my effort should be inline with the future of Xymon, as well as<br>
>> having maximal usefulness to others. It also needs to not break the<br>
>> existing heuristics used for the standard graphs (disk, inode,<br>
>> if_load, etc).<br>
>><br>
>> One idea is that I could add a "LINECOUNT" keyword into the graphs.cfg<br>
>> file, something like this:<br>
>><br>
>> [bindstats]<br>
>> TITLE DNS Queries and Responses<br>
>> YAXIS per second<br>
>> LINECOUNT 6<br>
>> ...<br>
>><br>
>> This would help with both data and status messages that are falling<br>
>> victim to this problem.<br>
>><br>
>> It also occurs to me that the multi-graph code doesn't seem (to me) to<br>
>> make sense for graph definitions that don't have @RRDIDX@ type indexed<br>
>> filename lookups. So perhaps the simplest thing would be to assume<br>
>> linecount is zero if FNPATTERN is not specified in the corresponding<br>
>> graph definition.<br>
>><br>
>> Other ideas?<br>
>><br>
>> Cheers<br>
>> Jeremy<br>
>> _______________________________________________<br>
>> Xymon mailing list<br>
>> <a href="mailto:Xymon@xymon.com">Xymon@xymon.com</a><br>
>> <a href="http://lists.xymon.com/mailman/listinfo/xymon" target="_blank">http://lists.xymon.com/mailman/listinfo/xymon</a><br>
</div></div></blockquote></div><br></div>