I'm trying to set up nutanix cluster on 3 Dell XC nodes using Dell customized Vmware 5.5 ISO (suggested in their documentation) and phoenix-3.0.1_NOS-22.214.171.124.iso
I was able to image the nodes: all the CVMs are up and running, they ping each other and the hypervisors, but when i try to manually create the cluster this error happens:
2015-12-18 03:58:49 WARNING genesis_utils.py:689 Failed to reach a node where Genesis is up. Retrying... (Hit Ctrl-C to abort).
Didn't find any documents speaking about "genesis_utils.py:689" so I tried generic solutions suggested for genesis troubleshooting. No one of them solved my problem.
Restarting genesis the genesis.out log reports the following output:
nutanix@:192.168.1.180:~$ tail -f ~/data/logs/genesis.out
2015-12-18 04:03:02 INFO server.py:113 GENESIS START
2015-12-18 04:03:02 INFO server.py:118 Factory config file is found
2015-12-18 04:03:02 INFO server.py:125 Starting the serve_http thread
2015-12-18 04:03:02 INFO node_manager.py:845 Assigning IP address 192.168.5.2 to eth1
2015-12-18 04:03:02 ERROR sudo.py:25 Failed to load file /var/run/dhclient-eth1.pid, ret 1, stdout , stderr cat: /var/run/dhclient-eth1.pid: No such file or directory
2015-12-18 04:03:03 INFO node_manager.py:845 Assigning IP address 192.168.5.254 to eth1:1
2015-12-18 04:03:03 ERROR sudo.py:25 Failed to load file /var/run/dhclient-eth1:1.pid, ret 1, stdout , stderr cat: /var/run/dhclient-eth1:1.pid: No such file or directory
2015-12-18 04:03:04 INFO node_manager.py:863 Link-local interface eth2 not created, skipping...
2015-12-18 04:03:05 ERROR esx.py:249 Failed to load the IP routing configuration for the ESX host, ret 255, stdout , stderr
2015-12-18 04:03:05 ERROR esx.py:220 Unable to load the route configuration for the ESX host
2015-12-18 04:03:05 CRITICAL decorators.py:46 Traceback (most recent call last):
File "/home/hudsonb/workspace/workspace/danube-126.96.36.199-stable-release/builds/build-danube-188.8.131.52-stable-release/python-tree/bdist.linux-x86_64/egg/util/misc/decorators.py", line 40, in wrapper
File "/home/hudsonb/workspace/workspace/danube-184.108.40.206-stable-release/builds/build-danube-220.127.116.11-stable-release/python-tree/bdist.linux-x86_64/egg/cluster/genesis/server.py", line 129, in run
File "/home/hudsonb/workspace/workspace/danube-18.104.22.168-stable-release/builds/build-danube-22.214.171.124-stable-release/python-tree/bdist.linux-x86_64/egg/cluster/genesis/node_manager.py", line 523, in initialize
File "/home/hudsonb/workspace/workspace/danube-126.96.36.199-stable-release/builds/build-danube-188.8.131.52-stable-release/python-tree/bdist.linux-x86_64/egg/cluster/host_upgrade_helper.py", line 437, in get_host_version
File "/home/hudsonb/workspace/workspace/danube-184.108.40.206-stable-release/builds/build-danube-220.127.116.11-stable-release/python-tree/bdist.linux-x86_64/egg/cluster/host_upgrade_helper.py", line 633, in get_host
AttributeError: 'NoneType' object has no attribute 'get_ipv4_config_object'
Do you have any suggestion?
Thanks in advance
Solved! Go to Solution.
By the looks of the trace seen in genesis logs.
It looks like genesis is unable to communicate with ESXi host.
Checked whether ESXi vmknic's are properly configured (Both public IP and Internal IP)
Internal IP of ESXi should be 192.168.5.1 and CVM's internal IP should be 192.168.5.2
thanks for your suggestion. I partially solved the problem launching the command fix_host_ssh.
I have no more messages like the previous one and i am able to see the status of the cluster via ssh (that is correctly unconfigured).
Remains the problem of cluster creation (CRITICAL cluster:756 Could not discover all nodes specified.)
Thanks for Suggestion Jon,
actually I am testing and studing Nutanix platform on Dell, it's not the case of a critical environment so every trouble is a new "topic" to study that i am happy to face with the help of the community.
Thanks again, it succeed!
Clearly the switch i connected the nodes is not performing very well, so the nodes need more time to communicate to each other, sure the cluster will proof me that... but at least i have a nutanix cluster :-)
Thanks a lot again!
No worries, happy to help out there. Just trying to make sure everyone gets the best support possible.
Secondarily, we recently released 4.5.1, so feel free to try out upgrades when you get a chance.
Olase Jon... could you help me about this problem ? I don't know what to do. I'm not able to create my cluster...
# cluster -s 192.168.1.160,192.168.1.161,192.168.1.162 c reate
option -s unrecognized - ignored
Usage: cluster <options> graphfile
-C k - generate no more than k clusters (0)
0 : no limit
-c k - use clustering method k (0)
0 : use modularity
1 : use modularity quality
-o <outfile> - output file (stdout)
-v - verbose mode
-? - print usage
I have checked again with my cluster and the "-s" option is currently being in used with "cluster" CVM command. It seems that your command has got something wrong.
Can you provide some info about your cluster environment?