Re: [Ganglia-general] Nodes dont show up in 2nd cluster.

[J Rangi]

Sorry forgot to paste the error.

What does this error mean, how can I fix this
When I start gmetad with debug level 10,

$ service gmetad restart
Shutting down GANGLIA gmetad:                              [  OK  ]
Starting GANGLIA gmetad: Going to run as user nobody
Sources are ...
Source: [SQL Cluster, step 15] has 1 sources
       192.168.240.145
xml listening on port 8651
interactive xml listening on port 8652
cleanup thread has been started
Data thread 89873328 is monitoring [SQL Cluster] data source
       192.168.240.145
[SQL Cluster] is an OLD version
hash_create size = 1024
hash->size is 1031
hash_create size = 50
hash->size is 53



I guess this should create "SQL Cluster" directory in /var/lib/ganglia/rrds/

Thank you,
-Jai

Ian Cunningham wrote:

J Rangi wrote:

Hello,
I am having trouble in configuring two separate clusters in Ganglia.

Version
ganglia-gmond-3.0.1-1
ganglia-monitor-core-lib-2.5.7-1
ganglia-gmetad-3.0.1-1
ganglia-web-3.0.2-1

/etc/gmetad.con  on web1

data_source "WEB SERVER" web1:864  Web2:8649 web3:8649 web3:8649
data_source "SQL Cluster" SQL1:8649 SQL2:8649 SQL3

/etc/gmond./conf on web1, web2, web3, sql1, sql2, sql3

globals {
 setuid = yes
 user = nobody
 cleanup_threshold = 300 /*secs */
}
cluster {
 name = "Web Cluster"
 owner = "Our Company Name"
}

udp_send_channel {
 host = Web1
 mcast_if = eth1
 port = 8649
}

udp_recv_channel {
 port = 8649
        acl {
          default = "allow"
          access {
            ip = 192.168.248.0
            mask = 32
            action = "allow"
          }
        }
}
tcp_accept_channel {
 port = 8649
acl{
 default="allow"
 access {
 ip= 192.168.240.0
 mask = 32
 action = "allow"
 }
}
}

gmond is running on all the nodes adn gmetad is running on web1 (master for ganglia).

Now the problem is that all the nodes are displayed in the same cluster.
data_source "WEB SERVER" web1:864  Web2:8649 web3:8649 web3:8649
data_source "SQL Cluster" SQL1:8649 SQL2:8649 SQL3:8649

From these statements I was expecting that web1, web2, adn web3 will be in WEB SERVER cluster and SQL1, SQL2 and SQL3 will be in SQL Cluster.

My SQL Cluster is empty.
Can some one give me hint what I am missing.
Thank you,
-Jai

Jai,

Try creating a second gmond.conf for the SQL Cluster. Set the udp_send host to sql1 instead of web1. sql1 will be your headnode for the SQL cluster. Then restart all the gmonds and gmetad.

You don't need to define all the nodes on all the clusters in the gmetad.conf, just the headnodes. The list is for failover, if the first headnode fails to respond, the second will be queried. In your current setup, you only have one headnode, web1.

Good Luck,
Ian