Re: [Ganglia-developers] ganglia.h cannot be compiled with C++
Hi, Bemard Thanks for the reply. Wrapping in C first, and use them will work for C++ application, but not ideal solution. On Tue, Apr 19, 2011 at 2:48 AM, Bernard Li bern...@vanhpc.org wrote: Hi Kazuki: On Mon, Apr 18, 2011 at 1:30 AM, Kazuki Ohta k...@preferred.jp wrote: I'm now trying to implement C++ server, with native Ganglia metrics support. The server throws to local gmond with its own metrics. [kzk@red0 tmp]$ ganglia-config --version Ganglia 3.1.7 However, ganglia.h cannot be compiled with g++, becuase of the typedef problem. https://gist.github.com/924982 Not sure if this is helpful to you, but we have some documentation on how to write gmond C modules (and Python modules) here: https://sourceforge.net/apps/trac/ganglia/wiki Good luck, Bernard -- -- Kazuki Ohta: http://kzk9.net/ -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers
Re: [Ganglia-developers] ganglia.h cannot be compiled with C++
Hi Kazuki: On Mon, Apr 18, 2011 at 11:52 PM, Kazuki Ohta k...@preferred.jp wrote: Thanks for the reply. Wrapping in C first, and use them will work for C++ application, but not ideal solution. The documentation I provided is for writing gmond C modules, however we do support writing modules in C++ at least according to the release notes: http://sourceforge.net/apps/trac/ganglia/wiki/ganglia_release_notes#GangliaRelease3.1.1 Perhaps you can use the guide and try writing a C++ module and see if you can make it work? I also found this thread, perhaps it will come in handy: http://www.mail-archive.com/ganglia-developers@lists.sourceforge.net/msg05278.html On the other hand, it looks like you're trying to write a C++ program using ganglia header files, that is not the same as writing a C++ gmond module. You need to figure out whether writing a module will work in your case. Good luck, Bernard -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers
[Ganglia-developers] let Grid and Cluster Load graphs at the web 2.0 interface lock like at the old interface
Attached is a patch to let the Grid and Cluster Load graphs lock like at the
old interface.
Add possibility to define a custom data source in the json template with DS
which will get used instead of the default 'sum'. To handle situations where
the data source is not available in all rrds define the context.
For example the number of Nodes:
data source 'num' only available in the summary rrds of 'cpu_num'
definition in the json template looks like:
{ metric: cpu_num, color: 00FF00, label: Nodes, line_width: 2,
type: line, DS: num, context:
{ cluster: true, meta: true } }
The metric will get ignored at the host context but not at the meta and
cluster context.
Markus
--
Markus Köberl
Graz University of Technology
Signal Processing and Speech Communication Laboratory
E-mail: markus.koeb...@tugraz.at
Index: graph.d/load_report.json
===
--- graph.d/load_report.json (revision 2574)
+++ graph.d/load_report.json (working copy)
@@ -5,6 +5,8 @@
vertical_label : Loads/Procs,
series : [
{ metric: load_one, color: BB, label: 1-min, type: stack },
+ { metric: cpu_num, color: 00FF00, label: Nodes, line_width: 2, type: line, DS: num, context: { cluster: true, meta: true } },
+ { metric: cpu_num, color: FF, label: CPUs, line_width: 2, type: line },
{ metric: proc_run, color: 2030F4, label: Procs, line_width: 2, type: line }
]
}
Index: functions.php
===
--- functions.php (revision 2574)
+++ functions.php (working copy)
@@ -870,7 +870,10 @@
// Loop through all the graph items
foreach( $graph_config[ 'series' ] as $index = $item ) {
-
+ // ignore item if context is not defined in json template
+ if ( $item[ 'context' ] and $item[ 'context' ][ $context ] == false )
+ continue;
+
$rrd_dir = $conf['rrds'] . / . $item['clustername'] . / . $item['hostname'];
$metric = sanitize( $item[ 'metric' ] );
@@ -884,8 +887,14 @@
$unique_id = a . $index;
$label = str_pad( sanitize( $item[ 'label' ] ), $max_label_length );
- $series .= DEF:'$unique_id'='$metric_file':'sum':AVERAGE ;
-
+
+ // use custom DS defined in json template (default = 'sum')
+ $DS = sanitize( $item[ 'DS' ] );
+ if($DS!='')
+ $series .= DEF:'$unique_id'='$metric_file':'$DS':AVERAGE ;
+ else
+ $series .= DEF:'$unique_id'='$metric_file':'sum':AVERAGE ;
+
// By default graph is a line graph
isset( $item['type']) ? $item_type = $item['type'] : $item_type = line;
--
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
Re: [Ganglia-developers] (no subject)
Read the documentations. On 19-Apr-2011, at 8:36 PM, aahan noor wrote: Hi i am new to ganglia. i need to know from where i can get help of my several questions regarding ganglia.my 1st question is that what is the task of gmond module which is written in c, how gmond config file and module is connected.? and 2nd question is that in linux operating system, in which directory gmond module is located? waiting for yours favour -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev___ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev___ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers
Re: [Ganglia-developers] let Grid and Cluster Load graphs at the web 2.0 interface lock like at the old interface
On Apr 19, 2011, at 9:59 AM, Markus Köberl wrote:
On Tuesday 19 April 2011, Alex Dean wrote:
On Apr 19, 2011, at 8:34 AM, Markus Köberl wrote:
{ metric: cpu_num, color: 00FF00, label: Nodes, line_width:
2, type: line, DS: num, context: { cluster: true, meta:
true } }
This seems like a good idea to me.
What about formatting like this:
{ metric: cpu_num, color: 00FF00, label: Nodes, line_width:
2, type: line, DS: num, contexts: [cluster, meta] }
You could implement with:
if ( isSet($item[ 'contexts' ]) array_key_exists($context,
$item['contexts']) ) { continue;
}
The context names seem like data here, not keys, so I just thought an array
might be more appropriate.
Yes, it is much simpler now.
Markus
Ah, yeah, especially since my suggested change implemented the opposite of your
idea. Oops. :)
Also, I'm remembering to reply to everyone this time.
alex
--
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev
___
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
