2nd node won't come up after initial database creation

Adam_2

I install the rpm, run the install_vertica script, create db through admintools. Everything goes without a hitch until the system tries to bring the 2nd node up. I then get a neverending:

        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)
        Node Status: v_compdw_node0001: (UP) v_compdw_node0002: (DOWN)

I get no warnings on pre-requisites and both servers are identical hardware. I've looked through a handful of logs but can't see anything that points to a definitive issue.

Here's the commands I ran:

[root@centosv1 ~]# rpm -Uvh /tmp/vertica-7.0.0-1.x86_64.RHEL5.rpm
[root@centosv1 ~]# /opt/vertica/sbin/install_vertica --hosts centosv1,centosv2 --rpm /tmp/vertica-7.0.0-1.x86_64.RHEL5.rpm --dba-user dbadmin --control-network 172.17.0.255 --point-to-point
[root@centosv1 ~]# su - dbadmin
[dbadmin@centosv1 ~]$ admintools

create database with all default except for catalog/data paths. Changed those to /home/catalog and /home/data

I had originally left off --control-network and --point-to-point but received other errors. I added --control-network after researching other community posts for an error. After that, I was getting to the same point I'm at now but node 1 would get stuck in initializing with node 2 down instead of node 1 up and node 2 down. Adding --point-to-point at least got node 1 up.

Not sure where to go from here

Nimmi_gupta

Can you attach the vspread.conf? You can find vspread.cong in catalog directory.

Adam_2

After some communication with support and referencing this post, https://community.vertica.com/vertica/topics/adding_2nd_node_to_vertica_7_0_1_not_working_first_node..., we came to the same conclusion. I had been avoiding that file in particular due to the 1st line in the file being "# Auto-generated by vertica - do not edit". There's apparently a bug that erroneously puts loopback entries in spread.conf. I manually updated the file based on our actual node IPs and the cluster is up and running. I'm surprised that this isn't a more widely reported problem on new 7.0 installs.

Prasanta_Pal

Not sure if you have resolved the problem, but I believe you have visited the vertica 7 documentation
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/InstallationGuide/InstallingVertica/RunTheInstallScript.htm%3FTocPath%3DInstallation%20Guide|Installing%20HP%20Vertica|Installing%20Using%20the%20Command%20Line|_____3

Run 'ifconfig' on the hosts and see if the subnet of control_network matches with the subnet of the hosts.
what are the IPs of your hosts

There should be a spread.config file in catalog directory, you may paste the content