You might want to consider summarizing the statistics for all the GPUs on the
node - this is the approach taken by sFlow agent:
http://sflow.org/sflow_nvml.txt
FYI Gmond natively understands the sFlow GPU metrics:
http://blog.sflow.com/2012/10/using-ganglia-to-monitor-gpu-performance.html
Actually, looking at the code, it looks the the message type is explicit, so
you wouldn’t need any configuration:
msg-version != EBT_ULOG_VERSION
On Mar 1, 2014, at 1:54 PM, Ben Hartshorne gang...@green.hartshorne.net wrote:
I've changed my mind about reorganizing the repo. The correct way to
Please ignore the non sequitur - I accidentally sent this reply to the wrong
email thread.
On Mar 1, 2014, at 2:30 PM, Peter Phaal peter.ph...@inmon.com wrote:
Actually, looking at the code, it looks the the message type is explicit, so
you wouldn’t need any configuration:
msg-version
Adrian,
The Host sFlow agent exports UUIDs (and is compatible with Ganglia 3.2+).
http://blog.sflow.com/2011/07/ganglia-32-released.html
The daemon relies on command utilities invoked at startup to find the UUID (for
reasons of portability that Dave mentioned). You could take a look at the
:04 AM, Peter Phaal peter.ph...@inmon.com wrote:
Simon,
Ganglia 3.2+ can collect and display sFlow metrics:
http://blog.sflow.com/2012/01/using-ganglia-to-monitor-virtual.html
You can monitor the virtual machines by installing Host sFlow agents
(http://host-sflow.sourceforge.net
definitely let you
know about my problems.
If you can't answer my previous question to whom should I send it?
Cheers,
On Wed, Jun 20, 2012 at 8:24 PM, Peter Phaal peter.ph...@inmon.com wrote:
I can't help you with your questions about gmetad, but I would be very
interested in hearing
:
On Wed, Jun 20, 2012 at 8:57 PM, Peter Phaal peter.ph...@inmon.com wrote:
There are people on this list who should be able to answer questions relating
to gmetad (although you might have better luck on the ganglia-general list).
My expertise is limited to sFlow.
Btw. sFlow can use TCP as well
Simon,
Ganglia 3.2+ can collect and display sFlow metrics:
http://blog.sflow.com/2012/01/using-ganglia-to-monitor-virtual.html
You can monitor the virtual machines by installing Host sFlow agents
(http://host-sflow.sourceforge.net/ ) on the hypervisors. The sFlow agents
export libvirt metrics
It would be very easy to eliminate the dependency on the JSON library entirely.
The JSON library is useful if you want import JSON, or export complex
structures. However, in this case where the only requirement is to export a
simple hash table the library isn't necessary. Exporting the gmond
I think a JSON option would make the gmond data more accessible. If the hood is
going to be open on the transport, an HTTP option would further open up access
to the data. HTTP would provide a way to pass query arguments and return
specific metrics, or metrics for a particular host, tag etc. I
Vladimir,
It's great to see the progress with Ganglia 4.0.0! Once comment on the release
notes, sFlow HTTP metrics are also supported (in addition to the Virtual
machine, Java VM and Memcache metrics that are listed).
FYI For anyone interested in monitoring web server farms, sFlow agents
You need to be running Ganglia 3.2 and send the sFlow to gmond (not gmetad).
The following article describes how to configure gmond to receive sFlow:
http://blog.sflow.com/2011/07/ganglia-32-released.html
On Jan 11, 2012, at 12:50 PM, Im Root wrote:
Vlad,
I downloaded this agent and
If there is interest, I would be happy to contribute a section on the sFlow
functionality added in Ganglia 3.2
Peter
On Dec 7, 2011, at 10:04 AM, Matt Massie wrote:
Are there any more volunteers?
Later today, I'm going to submit the list of volunteers to O'reilly and start
getting
Bernard,
Thanks for committing the patch. Hopefully more people will try the sFlow
functionality and we can get additional feedback.
On Oct 26, 2010, at 12:47 AM, Bernard Li wrote:
If I could make just one comment: right now if I run gmond in debug
mode, it shows that the sFlow metrics are
Brad,
Thanks for the information on spoofing in modules.
The way that I would envision an sflow module working would be similar to the
spoofing example module that is currently checked into the Ganglia SVN
repository. The spoofing module can be found at
Brad,
Thanks for the feedback. My comments are in-line.
On Oct 7, 2010, at 12:27 PM, Brad Nicholes wrote:
Sorry to jump into this thread so late but I thought that I would throw my 2
cents in.
I finally got a chance to take a look at the code. I was able to compile it
but ran into
, we will send a patch containing the changes to
gmond.c.
Peter
-Original Message-
From: Bernard Li [mailto:bern...@vanhpc.org]
Sent: Monday, August 09, 2010 10:04 PM
To: Peter Phaal
Cc: ganglia-developers@lists.sourceforge.net
Subject: Re: [Ganglia-developers] sFlow counters in Ganglia
Hi
I wanted to see if there was any interest in the Ganglia community in providing
a way to incorporate sFlow counters in Ganglia. Looking at the source code it
looks like it would be relatively straightforward to include modify gmond to
allow an sflow_recv_channel (basically a udp_recv_channel
metrics.
Some preliminary results have been published:
http://blog.sflow.com/search/label/server
Peter
On Mar 12, 2010, at 7:28 AM, Peter Phaal wrote:
It should be possible integrate both approached in the library. Rate
calculations are a layer on top of the raw counters. One approach would
the Ganglia project. We may create a similar library for our
project, exposing the underlying counters.
Peter
On Mar 7, 2010, at 10:57 AM, Daniel Pocock wrote:
Peter Phaal wrote:
Hi all,
We are considering using libmetrics as a stand-alone library. In its current
form, with the separate
On Mar 7, 2010, at 10:57 AM, Daniel Pocock wrote:
Peter Phaal wrote:
Hi all,
We are considering using libmetrics as a stand-alone library. In its current
form, with the separate configure script, it looks like it would ideally
suite our purposes. I noticed a recent discussion about
Hi all,
We are considering using libmetrics as a stand-alone library. In its current
form, with the separate configure script, it looks like it would ideally suite
our purposes. I noticed a recent discussion about dropping the separate
configure script[1] and wanted to express our interest in
22 matches
Mail list logo