Support Center > Search Results > SecureKnowledge Details
RouteD daemon might consume CPU at very high level on ClusterXL member running Gaia OS, when there are issues with cluster sync interfaces
Symptoms
  • Output of 'top' command on ClusterXL member shows that 'routed' daemon consumes CPU at very high level (which might cause fail-over).

  • /var/log/messages file on problematic on ClusterXL member repeatedly shows:

    • routed[PID]: cpcl_instance_manager_listen_accept(N): task_accept failed, errno = 24
    • routed[PID]: cpcl_instance_manager_listen_accept(N): peer_alloc failed, errno = 0
  • /var/log/routed.log file on problematic on ClusterXL member under trace ('Global - Cluster') repeatedly shows:

    cpcl_slave_restart(N): Entering...
    slave: quiting peer (id = 1).
    peer_remove(N): Entering !!!!
    peer_remove(N): resetting sockets
    cpcl_slave_quit(N): closing peer socket 25
    slave: closing slave sock (id = 25).
    cpcl_clear_recv_queue(N): entering
    entering cpcl_slave_init()
    cpcl_slave_init(N): instance 0 connection waiting for select to return 
    cpcl_slave_restart(N): Exiting...
    cpcl_slave_connect_complete(N): entering
    cpcl_slave_connected_to_master(N): entering
    cprd: state switching from 1 to 1
    cpcl_slave_connect_complete(N): exiting
    
Cause

In some scenarios, if there are issues with cluster synchronization interfaces (e.g., when 2 physical Sync interfaces are defined, which is against sk92804), RouteD might not be able to accept a connection from peer cluster member. Due to a socket not being closed, eventually RouteD reaches the limit of open files for a process.


Solution
Note: To view this solution you need to Sign In .