In Management Console one node of three has "unknown" info but Status is UP

We restarted the problematic agent and still  netstat -a | grep 5444 didn't show anything.
In /opt/vertica/log/agent.log we saw the following :
WARNING:pyinotify:Unable to retrieve Watch object associated to <_RawEvent cookie=0 mask=0x2 name=ExecutionEngineProfiles_462377823091961.log wd=2 


Comments

  • Hi!

    https://community.vertica.com/vertica/topics/the_problem_with_mc
    There is a defect on Vertica and MC 7.01 that when MC and vertica clusters are on different networks, and if user imported the cluster using public IP of the cluster, Management Concole may not work correctly.  Is this your case?  This issue has been fix in 7.1.

    If you install management console, as a test, on one of the node of the clusters themselves (one MC per cluster), will you have same issue?

     

    Regards,

    Han

  • Hi,
    I work with leonidn. the MC is installed on Node1. When I open the MC and input the Node1 IP for it to connect to, I get an error. If I enter any other node IP, I connect ok and then I see Node1 as "UNKNOWN".
  • You can try starting the agent from the agent script /opt/vertica/agent/agent.sh and see if that will bring up the 5444 port on the node.
    /bin/bash /opt/vertica/agent/agent.sh /opt/vertica/config/users/dbadmin/agent.conf &

     
  • Thanks. I'll try it.
  • Hi Raul,
    It didn't help. Still same issue.
  • Can you send the output of the following -
    rpm -qa | grep vertica
  • vertica-7.0.0-1.x86_64
    vertica-console-7.0.0-0.x86_64

  • I am having the same issue. I have the management console installed on Node 1 and Node 2 is showing an unknown status for mem, cpu, and host name. Also when going into management and double clicking the node for more detail I get the following error message, "Connection to the agent is lost. Logs are unavailable."

    I am running version 7.1

    # rpm -qa | grep vertica
    vertica-R-lang-7.1.1-0.x86_64
    vertica-7.1.1-0.x86_64
    vertica-console-7.1.1-0.x86_64

    Update**
    Actually running the following command Raul mentioned did the trick for me.
    /bin/bash /opt/vertica/agent/agent.sh /opt/vertica/config/users/dbadmin/agent.conf &

  • We finally found the problem. We tried to start simply_fast.py (as it was using port 5444 on the other nodes) and got this error:
    error: [('pem routines', 'pem_read_bio', 'no start line'), ('ssl routines', 'ssl_ctx_use_privatekey_file', 'pem lib')] vertica agent terminated.
    We checked what the python code was looking for and we saw that it looks for agent.pem. I don't know why yet but the file existed on the node, even though we don't use ssl. we removed the agent.pem file and restarted the agent. problem fixed.
    Thanks for your help! 

Leave a Comment

BoldItalicStrikethroughOrdered listUnordered list
Emoji
Image
Align leftAlign centerAlign rightToggle HTML viewToggle full pageToggle lights
Drop image/file