. Does anyone
running into the same problem?
Thanks,
paul
--
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo
but maybe missing something I did not see.
Thanks for the quick response.
paul
On Jun 5, 2015, at 9:58 AM, Jesse Becker haw...@gmail.com wrote:
goon?
Not familar with that, except in hockey and nursery rhymes[1].
Gmetad is responsible for creating the RRD files. Make sure the value
verbose, and probably not helpful right
now.
Also run nc against the gmond collector, to make sure it has the
metrics it should.
On Fri, Jun 5, 2015 at 11:40 AM, Paul p...@space.mit.edu
mailto:p...@space.mit.edu wrote:
Hi Jesse,
Here are the output from gmod and gmetad. x.x.x.xs are all
Sorry, I was not thinking.
nc localhost 8651 did give me full of the messages for all collector and nodes,
include:
cluster info, node info, IP, REPORTED#, locarion, TN, TMax, DMax,
GMOND_STARTED, TAGS=“”
paul
On Jun 5, 2015, at 12:11 PM, Paul p...@space.mit.edu wrote:
running nc
Hi Jesse,
Here are the output from gmod and gmetad. x.x.x.xs are all correct IPs. I know
gmond works right. I ran tcpdump and got correct connections and ports. I am
not sure the output from gmetad is normal or missing any piece.
Thanks,
paul
# gmond -d 5
loaded module: core_metrics
loaded
.
Anyone have the settings in a similar way?
Thank you!
Paul
--
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all
Thank you! Jay.
That is what I want to know. You answered my question. I will do it differently.
cheers,
paul
On Feb 8, 2015, at 11:36 PM, jayadevan Chembakassery jayadev...@gmail.com
wrote:
Ok.
The Tree being-- GRID - CLUSTERS- NODES , what you expect to see
in the Choose
gmond2 port2 */
data_source Cluster2 gmonhost:port3/* from gmond2 port3 */
in gmond.conf, you can specify multiple ports for sending and receiving. I
thought I can use each port to handle a group of nodes.
Thank you!
paul
On Feb 8, 2015, at 6:22 PM, jayadevan Chembakassery jayadev...@gmail.com
Hi Waleed
I recently wrote a python LSF module for my last contract. It reported
metrics on the jobs submitted to LSF as opposed to
monitoring LSF itself (sbatchd,lim,res etc).
Is this what you want?
If so I could ask if the module could be made available
Regards
Paul
On 11 December 2012
Thank you Jochen
--
Paul Hewlett
Write documentation as if whoever reads it is a violent psychopath who knows
where you live.
Steve English, as quoted by Peter Langston
http://www.quotegarden.com/programming.html
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223 405923
skype
--
Paul Hewlett
Write documentation as if whoever reads it is a violent psychopath who knows
where you live.
Steve English, as quoted by Peter Langston
http://www.quotegarden.com/programming.html
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223 405923
skype: paul-at-arm
Hi Guys
I have just installed ganglia-web 3.5.2 and noticed that on the main web page
if I click on 'Same' then all my graphs are autoscaled. If I click on Auto then
all the graphs are scaled the same way.
Surely this is the wrong way round? (or am I stupid...)
Regards
--
Paul Hewlett
by the length of the time period and
multiply by 100.0 and by JIFFIES_PER_SEC
Regards
--
Paul Hewlett X25250
http://www.theregister.co.uk/2012/06/25/rbs_natwest_what_went_wrong/
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223 405923
skype: paul-at-arm
www.arm.com
End of Ganglia
a plain 'sdb' instead of 'sdb2' or modify the diskstat.py script in
/usr/lib64/ganglia/python-modules
Regards
--
Paul Hewlett X25250
http://www.theregister.co.uk/2012/06/25/rbs_natwest_what_went_wrong/
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223 405923
skype: paul-at-arm
way of submitting them?
Either universal diffs or should I check out the git tree?
Regards
--
Paul Hewlett X25250
http://www.theregister.co.uk/2012/06/25/rbs_natwest_what_went_wrong/
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223 405923
skype: paul-at-arm
www.arm.com
4. Re
= 1
time_threshold = 30
metric {
name_match = procstat_(.+)_cpu
}
metric {
name_match = procstat_(.+)_mem
}
}
Regards
--
Paul Hewlett X25250
http://www.theregister.co.uk/2012/06/25/rbs_natwest_what_went_wrong/
ARM Ltd
110 Fulbourn Road, Cambridge, CB1 9NJ
Tel: +44 (0)1223
,autoconf
are specified and the 'bootstrap' script mentioned has been removed -(it was in
3.3.1 and I used it successfully)
I would really appreciate some help or suggestions -
Regards
--
Paul Hewlett X25250
http://www.theregister.co.uk/2012/06/25/rbs_natwest_what_went_wrong/
ARM Ltd
110 Fulbourn
I bollixed up the information stored in at least one __SummaryInfo_
directory. Is there any way to explicitly force a rebuild of a
particular __SummaryInfo__ directory structure from the lower level data
that appears to be intact?
-- Thanks,
--Paul Townsend (ITaP-ITSO-RCAC-Sysadmin _AT_
it
partially opens and then immediately closes again.
Thanks,
Paul
--
/***/
Paul L. Peltz, Jr.
University of Tennessee, Knoxville
Electrical Engineering and Computer Science Dept. - ICL
IT Administrator II
865-974-1075
http
Thanks, that worked perfectly.
Paul
On Thu, Aug 25, 2011 at 10:55 AM, Peter Piela peter.pi...@terascala.comwrote:
Paul,
Attached is a host_view.tpl that has a fix for the issue. It goes in the
templates/default directory. Please send me a note to confirm the fix. I
will then commit
I'd like to also add that if you use unicast with multiple udp_send_channel,
you will want to restart your gmonds first. Stop both gmonds, then start them
up. Then restart gmetad.
-Paul
From: Bernard Li bern...@vanhpc.orgmailto:bern...@vanhpc.org
Date: Mon, 2 May 2011 10:14:29 -0700
To: Lim
.
Paul
This e-mail and any files transmitted with it are
confidential and may be protected by legal privilege.
If you are not the intended recipient, please notify
the sender and delete the e-mail from your system.
This e-mail has been scanned for malicious content but
the internet is inherently
Is the reverse lookup for the host setup correctly? That's probably the most
common cause. Or check for any errors in /etc/hosts.
-Paul
From: Ivan Pulido [mailto:mefistofele...@gmail.com]
Sent: Wednesday, October 13, 2010 4:31 PM
To: ganglia-general@lists.sourceforge.net
Subject: [Ganglia
to the file or something like that.
I tried using nohup option, but with no luck.
Any ideas?
Thanks
Paul
--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field
Thanks Bernard
Again as before It worked when I tested it in terminal using cli, but when I
placed it in cron, the output was trucked again.
any other ideas?
-Paul
On Thu, Jan 14, 2010 at 11:18 AM, Bernard Li bern...@vanhpc.org wrote:
Hi Paul:
Have you considered using netcat instead
but I'll wait a few days to see how this current bit of
lash up survives.
Thanks again for all your help, it's much appreciated.
Cheers,
Paul
--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event
which take me a while a longer...
Cheers,
Paul
--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your
/show_bug.cgi?id=48029
and it is apparently fixed in the next version 1.3.10. When it is released
I'll rebuild and post back here confirming whether it has been fixed.
Thanks all for help and credit to Carlo Marcelo Arenas Belon for the fix!
Cheers,
Paul
On Tue, 27 Oct 2009, Paul Sobey wrote:
Maybe
Sounds like you might be seeing the same bug I was. Can be worked around
by rebuilding apr. See here:
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=244
Good luck!
Paul
On Tue, 17 Nov 2009, Rick Cobb wrote:
Yes. We would see the traffic on other machines, but we would
the python module to watch
the 200 zfs filesystems on one of my thumpers!
Cheers,
Paul
--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend
Just some further comments on Paul's case:
- I note Paul is using gcc, whereas I'm building and testing with Sun Studio
on the OpenCSW build farm - Sun's compiler is now a free download, and it is
used to build all the CSW libraries (including those used by Ganglia), so
this is now
On Wed, Oct 28, 2009 at 4:37 AM, Paul Sobey bud...@the-annexe.net wrote:
We try to build most things internally. There are good reasons which I'll
gladly discuss with you off-list if you'd like, but suffice to say it would
be useful to be able to compile Ganglia ourselves. I have no idea how
I note from the Makefile Daniel posted:
# Depends: some issues exist getting the Python support working on
Solaris,
# Ganglia's configure.in needs to be further enhanced for this to work
I'll build with python support disabled for now then :)
Paul
On Wed, 28 Oct 2009, Bernard Li wrote:
Hi
welcome!
Paul
On Tue, 27 Oct 2009, Bernard Li wrote:
Hi Paul:
Have you tried these OpenCSW packages:
http://mirror.opencsw.org/testing.html
Or is there a specific reason why you would want to build Ganglia yourself?
Thanks,
Bernard
On Tue, Oct 27, 2009 at 2:52 AM, Paul Sobey bud
how I can get it to build?
Cheers,
Paul
On Mon, 26 Oct 2009, Bernard Li wrote:
Dear all:
Ganglia 3.1.4 is ready for testing at:
http://ganglia.info/testing/
The RPMs were built on CentOS x86. To build for your specific
platform, please use:
rpmbuild -tb ganglia-3.1.4.tar.gz
between 10 mins - one hour. We've tried compiling apr with the
--disable-nonportable-atomics just in that made a difference but the
result is the same.
Hmmm perhaps I should file that bug today!
Paul
--
Come build
Maybe you guys can file a bug at bugzilla.ganglia.info in case this
hasn't already been filed.
Done:
http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=244
let me know how I can help from here.
Cheers,
Paul
For Aix look at the following document written by Nigel Griffiths of nmon fame
:-
http://www.ibm.com/developerworks/wikis/display/WikiPtype/ganglia
From: Kyrre Førli [mailto:[EMAIL PROTECTED]
Sent: 10 December 2008 13:50
To:
is for the gmond instance(s) that'll receive info from
other gmonds in the same cluster. Then you can configure gmetad to poll
gmond in foo02 and foo02 (in this case).
-Paul
Hardik Shah wrote:
Hi Jesse,
Thanks for your suggestion!! But please let me know how can I change all
compute nodes
if HOSTNAME= and IP= are correct for
each host.
-Paul
Craig Simpson wrote:
Have gmond running on 2 hosts. But I can only gather from one at a
time, even with gmond running on both.
The server running gmetad is configured like this:
/etc/ganglia/gmetad.conf
data_source Conquest oralnx51
host01 and host02. (IMPORTANT!)
Basically, restart all gmond master(s) for the cluster.
2. restart gmetad
3. Then you can do a XML dump of gmetad using netcat as Bernard advised.
Grep for the gmetric items to see if they're gone.
4. If #3 is satisfactory, go and delete the rrds.
-Paul
[EMAIL
, so packages were already
available. I did try to build one with gmetad, but I gave up because it
was a bit of a pain. If you are familiar with Linux, I'd maybe run
gmetad with Linux and have gmond run on whatever flavor of *nix.
-Paul
Big Woobie wrote:
Does anyone have a single package
similar to yours. Although I rsync to disk every
hour and not every 5 minutes.
-Paul Choi
Ben Rockwood wrote:
I posted to the list some time ago about problems scaling Gmetad. I
found that after passing some number of monitored nodes I was showing
nodes failed.
I've finally worked out
I am assuming that you successfully built gmond and did make install.
Can you try running gmond (wherever it's installed) with --help?
Look under -c, --conf=STRING section and see where the default
gmond.conf should be.
$ sudo /usr/sbin/gmond --help
gmond 3.0.7
Purpose:
The Ganglia
I use iostat -x because I need extended stats. I'm interested in
measuring await and svctm provided by iostat -x to monitor the latency
on the disk device I'm monitoring.
-Paul
Bernard Li wrote:
Hi guys:
Thanks for the responses.
Both Paul and John mentioned iostat with -x -- if I am
Doug Nordwall wrote:
So, I've run into a situation where there is quite a bit of data
sitting on a lights out card that I'd like to get access to with
ganglia. Since it supports IPMI, it seems like a great option, and I'd
just use ipmitool. Every node has a lights out card, including an
.
Is there something I'm missing or something I've misconfigured?
I'm running Ganglia 3.0.5.
Thanks for your help.
Paul Choi
[EMAIL PROTECTED]
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual
Gah, sorry. Nevermind.
Turns out that su01's iptables needed to be modified. Should've checked
for the obvious first.
Sorry for the extra mail.
-Paul Choi
Plaxo, Inc.
Paul Choi wrote:
Hello,
I hope someone can shed a light on this.
I have a cluster called su.
Gmetad.conf pulls data from
is 200.
And Ganglia frontend reports cpu_speed as 2000Mhz.
I'm running gmond v3.0.5.
-Paul Choi
On Thu, 2008-01-24 at 12:04 -0600, [EMAIL PROTECTED] wrote:
Hi!
This is the error when running gmond in debug mode
[root at master samba]# gmond -d3
slurpfile() open() error on file
/sys/devices
Well, Richard, you got me there. You certainly have delved into it much
more than I have.
I think your explanation also explains why sometimes I see some host go
down and come back up randomly.
-Paul
-Original Message-
From: richard grevis [mailto:[EMAIL PROTECTED]
Sent: Thursday
inet addr:10.1.3.3 Bcast:10.1.255.255 Mask:255.255.0.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
Interrupt:24
Any ideas?
Thanks!
-Paul Choi
-
This SF.net email is sponsored
of the master gmond host, I see
that the hostnames have already been resolved. I take it that this host
has done the name resolving (reverse lookup).
I can't use route add -net ip dev interface trick since I'm using
a virtual interface (eth0:0) and I want to distinguish from eth0's IP...
-Paul
}
udp_send_channel {
mcast_join = bes02.plaxo.com
port = 8649
}
udp_recv_channel {
port = 8649
family = inet4
}
... Snip ...
Thanks for your help!
-Paul Choi
[EMAIL PROTECTED]
-
This SF.net email is sponsored by: Microsoft
Do you think it is worth posting examples of the gmetad.conf and gmond.conf
that we have working so that people out there can get up and going fast?
Personally I often find that looking at examples of configuration files is
the best way of learning.
Cheers
are actually double.
Eg: One cluster has 5 machines, it displays “Hosts up: 5” and “10 cpu
total”
Any clue why this strange behavior.
Is it becoz of hyperthreading, coz it shows up to linux as cpu0 and cpu1..
Thanks,
Vineet Agarwal
--
Paul Greidanus
CAD Administrator / Systems
The bandwidth loading issue is due to the default behavior of ganglia
nodes, which is to have every node a listener as well as a broadcaster.
In effect, every node is listening to every other node, and attempting
to broadcast all its known information to the gmetad server.
We have cut down
to the web server.
Paul
Rick Mohr wrote:
On Tue, 7 Jun 2005, Tom Van Overbeke wrote:
I found ganglia, and at first sight, it looks very much like what I need
(nice graphs that give an overview of a specific monitor on all the
servers). But from looking at the web site, it looks very much
do it anyway...
Paul
Matt Klaric wrote:
Changing to unicast as described below fixed the problem.
Any ideas why multicast isn't working? I know multicast is compiled
into the kernel and ifconfig shows that my interface is capable of it.
Thanks for all the help.
--Matt
On Mon, 2005
The way I would do it is this:
define only one data source
in the /etc/gmond.conf on the four systems that are not data sources,
set mute = no and deaf = yes in the global variables section, i.e.:
/* global variables */
globals {
mute = no
deaf = yes
debug_level = 0
setuid = yes
What you need to do is specify one from each cluster, maybe a few more
for redundancy, but you don't need each and every one. Everything with
common metad ports will broadcast to the rest of it's cluster, and you
need to get the data from only one of those machines..
Foster, Scott (MS)
if one node dies or is moved, then you
don't have to restart gmond on every single node to get it to 'forget'
the node... you just need to do it on the two listening nodes. Also,
network traffic is significantly reduced.
Paul
Princeton Plasma Physics Lab
Johnston Michael J Contr AFRL/DES wrote
broadcasting and listening at the same time. I don't understand
the actual mechanics, but they are now happy (or should I say
marginally happier... they never seem to really be happy ;-)
Paul
Princeton Plasma Physics Lab
Bernard Li wrote:
Hey Paul:
But I guess if the odd chance of both
on
all the nodes as I had to do when they all listened.
Thanks,
Paul
Matt Massie wrote:
paul-
i'm a little confused here. if you run all your cluster hosts in deaf
mode except for two hosts, then the amount of multicast traffic would
not change.
with your current configuration every host
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
Paul Henderson
UNIX Systems Engineering Group
Princeton Plasma Physics Laboratory
Princeton, NJ 08543
(609) 243-2412
.
I've checked all 192 gmond.conf files, and they each show that the
respective node is in the proper cluster. The gmetad.conf file on
'server' shows the two data_source lines properly.
Any solutions? Thanks for any help.
Paul
--
Paul Henderson
UNIX Systems Engineering Group
Princeton Plasma
I'm running 2.3.1b1 on Solaris 8, so I guess I'll hold off upgrading..
Once you do get it going you'll discover that, in fact, no metrics are
being reported because gmond isn't collecting any for Solaris yet. It
turns out (see ./gmond/machines/solaris.c) that the metric-specific
functions are
- How often does gmond sample core metrics of the host it's running on?
- In /etc/gmond.conf, is num_custom_metrics a hard limit? What happens
if this is set to, say 10, and I report 12 metrics?
- In 'ganglia-rrd.pl', when an RRD database is updated for some host
metric, is it updated with the
67 matches
Mail list logo