Metrics are not received by gmond

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Start gmond (Ubuntu: /etc/init.d/ganglia-monitor start)
2. sudo tcpdump -X -i lo udp dst port 8649
3. Run ganglia.gmetric.GMetric main() method
4. telnet localhost 8649

What is the expected output? What do you see instead?

telnet should output BOILINGPOINT test metric, but it is not there.
However, tcpdump shows the packets were sent.

What version of the product are you using? On what operating system?

jmxetric SVN revision 36 (latest)
Ubuntu 9.04
Sun JDK 1.6.0_16
ganglia-monitor 2.5.7-5

Please provide any additional information below.

GMetricTest fails because GMetricResult.getGMetric(name) returns null.

Original issue reported on code.google.com by nicolas....@gmail.com on 23 Sep 2009 at 6:39

[GoogleCodeExporter]

Hi Nicolas
I haven't tested it any earlier than Ganglia 3.0.7.

On Ganglia 3.0.7, I get:

19:00:11.934637 IP localhost.localdomain.45900 > localhost.localdomain.8649:
UDP, length 64
        0x0000:  4500 005c 0000 4000 4011 3c8f 7f00 0001  E..\..@.@.<.....
        0x0010:  7f00 0001 b34c 21c9 0048 fe5b 0000 0000  .....L!..H.[....
        0x0020:  0000 0006 7374 7269 6e67 0000 0000 000c  ....string......
        0x0030:  424f 494c 494e 4750 4f49 4e54 0000 0003  BOILINGPOINT....
        0x0040:  3130 3000 0000 0007 4345 4c53 4955 5300  100.....CELSIUS.
        0x0050:  0000                                     ..

$ telnet 0 8649 | grep BOIL
<METRIC NAME="BOILINGPOINT" VAL="100" TYPE="string" UNITS="CELSIUS" TN="0"
TMAX="0" DMAX="0" SLOPE="both" SOURCE="gmetric"/>

Regards
Jasper

Original comment by jasper...@gmail.com on 24 Sep 2009 at 5:07

[GoogleCodeExporter]

I have upgraded to ganglia-monitor 3.1.2-2
(http://packages.ubuntu.com/karmic/ganglia-monitor) and I'm still not seing the 
metric.

I have modified GMetric.main to set ganglia311 to true.

I'm seing two packets: (they differ slightly from yours)

17:04:44.655331 IP localhost.47049 > localhost.8649: UDP, length 180
    0x0000:  4500 00d0 0000 4000 4011 3c1b 7f00 0001  E.....@.@.<.....
    0x0010:  7f00 0001 b7c9 21c9 00bc fecf 0000 0080  ......!.........
    0x0020:  0000 000e 6e69 636f 6c61 732d 6275 7265  ....nicolas-bure
    0x0030:  6175 0000 0000 000c 424f 494c 494e 4750  au......BOILINGP
    0x0040:  4f49 4e54 0000 0000 0000 0006            OINT........
17:04:44.656105 IP localhost.47049 > localhost.8649: UDP, length 60
    0x0000:  4500 0058 0000 4000 4011 3c93 7f00 0001  E..X..@.@.<.....
    0x0010:  7f00 0001 b7c9 21c9 0044 fe57 0000 0085  ......!..D.W....
    0x0020:  0000 000e 6e69 636f 6c61 732d 6275 7265  ....nicolas-bure
    0x0030:  6175 0000 0000 000c 424f 494c 494e 4750  au......BOILINGP
    0x0040:  4f49 4e54 0000 0000 0000 0002            OINT........

Could it be a problem with my gmond config? I'm attaching it.

Original comment by nicolas....@gmail.com on 24 Sep 2009 at 9:06

[GoogleCodeExporter]

Attached /etc/ganglia/gmond.conf

Original comment by nicolas....@gmail.com on 24 Sep 2009 at 9:07

Attachments:

• gmond.conf

[GoogleCodeExporter]

I think you might be sending and receiving on different multicast networks. 
GMetric.main sends on localhost/8649 but your gmond is set to 
udp_recv_channel {
  mcast_join = 239.2.11.71
  port = 8649
  bind = 239.2.11.71
}
I think it should be more like:
udp_recv_channel {
  bind = 127.0.0.1
  port = 8649
}
Regards
Jasper

Original comment by jasper...@gmail.com on 29 Sep 2009 at 5:21

[GoogleCodeExporter]

That made it. Thanks!

Doc says I can add as many udp_recv_channel as I want, to I simply added the 
one you
suggested.

This issue can now be closed. It will be useful for others that may encounter 
the
same problem.

Original comment by nicolas....@gmail.com on 29 Sep 2009 at 4:38

[GoogleCodeExporter]

Original comment by jasper...@gmail.com on 29 Sep 2009 at 5:28

• Changed state: WontFix