Automic Workload Automation

  • Private Community
Back to discussions
Expand all | Collapse all

New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

  • 1.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 04:23 PM
    One of my users updated the OS on their agent.  I've already fixed one error they had in their config file on their first restart try (system name) but can't see the obvious answer for this one.  On restart, it is giving an Invalid bind call.  Other than the messages documentation, I can't find anything to tell me how to correct/troubleshoot the error.   Can anyone point me to the fix for this?  Thanks.

    20180129/135749.898 - U02003054 The Agent was able to gain user privileges of real user 0.
    20180129/135749.900 - U02000020 Environment: Hardware = 'x86_64/8'.
    20180129/135749.900 - U02000021 Environment: Software = 'Linux'.
    20180129/135749.901 - U02000022 Environment: SW version = '3.10.0-693.17.1.el7.x86_64'.
    20180129/135749.901 - U02000093 Environment: program addressing mode = '64-bit'.
    20180129/135749.903 - U02000045 Maximum number of open file descriptors a process may have: '32768'
    20180129/135749.907 - U02003006 Invalid 'bind' call, error: ('98' - 'Address already in use'). New attempt in 10 seconds.
    20180129/135759.908 - U02003006 Invalid 'bind' call, error: ('98' - 'Address already in use'). New attempt in 10 seconds.
    20180129/135809.908 - U02003006 Invalid 'bind' call, error: ('98' - 'Address already in use'). New attempt in 10 seconds.
    20180129/135819.908 - U02003001 Port number '2300' is being used. Canceling Agent.
    20180129/135819.909 - U02000002 Agent 'HSGCCL956' version '11.2.1+build.390' ended abnormally.
    20180129/135819.909 - U02000041 Shutdown Agent 'HSGCCL956'.
    20180129/135819.910 - U02000002 Agent 'HSGCCL956' version '11.2.1+build.390' ended abnormally.


  • 2.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').
    Best Answer

    Posted Jan 29, 2018 04:43 PM

    It says "20180129/135819.908 - U02003001 Port number '2300' is being used. Canceling Agent"

    Please check for any processes running which shouldn't be running on that port or try providing a different port number in the .ini file. That should help.



  • 3.  RE: New OS on Agent - **** 'bind' call, error: ('98' - 'Address already in use').

    Posted Nov 16, 2021 08:32 AM
    Hi is there any way to find out which port num should I used in ini file so that agent is installed successfully please help me out

    Original Message


  • 4.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 04:48 PM
    TimOsgood611330

    You can use the command below if it is installed to find the process using the port 2300.  Usually, root user has it:

    lsof  -i :2300 



    Thanks
    Bob


  • 5.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 04:59 PM
    Shravani.  thanks for the suggestion.  I put the entire queue for this agent on hold as there were 14 jobs active, waiting for host or waiting for predecessor.  Still get the same error.  Do I need to actually stop all tasks in the queue including the schedule?   Due to our firewall rules port 2300 is the only one available.


  • 6.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 05:14 PM
    Thanks Bob.  I passed that on to the agent admin. waiting for results.

    The results are:
    [root@hsgccl956 bin]# lsof  -i :2300
    COMMAND   PID USER   FD   TYPE  DEVICE SIZE/OFF NODE NAME
    ucxjlx6 25220 root    7u  IPv4 2414763      0t0  TCP *:cvmmon (LISTEN)

    Not sure what that's telling me.



  • 7.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 05:45 PM
    Is it possible that the system believes 2300 is in use, even thought it isn't?  The MQ tables on the DB have connection information.  Usually MQSRV will have the connected agents and server processes.  If you have access to the SEC_SQLI object you can run a quick SELECT * ON MQSRV and see what the system thinks is connected.


  • 8.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 29, 2018 08:19 PM
    TimOsgood611330

    please execute the command below and see the started date. If date is old, then you may need to kill it.

    ps -ef|grep 25220

    Do you have other agent installed on the same server ? if not, I suggest that you kill it. Then, try to restart the agent again and see if the issue remains.


    Thanks
    Bob


  • 9.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 30, 2018 04:14 AM
    "lsof" shows you open files. While that may include virtual files for ports, it's probably better to use netstat:

    netstat -anp | grep 2300

    Output will look like this:

    # netstat -anp | grep 2300
    tcp        0      0 0.0.0.0:2300                0.0.0.0:*                   LISTEN      2661/ucxjlx6

    2661 is the PID of the process occupying the port, you can use "kill <PID" to kill it, "kill -9 <PID>" if it refuses to die otherwise.

    After killing it, it may take some time for the port to be usable again. This is because of something called TIME_WAIT. The TIME_WAIT state in TCP/IP "represents waiting for enough time to pass to be sure the remote TCP received the acknowledgment of its connection termination request." You may need to wait up to two minutes in Linux for the port to be accessible again.




  • 10.  New OS on Agent - Invalid 'bind' call, error: ('98' - 'Address already in use').

    Posted Jan 31, 2018 11:43 AM
    Thank you all for your responses, it will help if this issue arises legitimately.  However, I finally got to see their ini file that was causing the errors.  Apparently after the upgrade to production, they somehow put an old dev ini file on prod and that was the root cause of the issues.  I fixed all the appropriate entries and now the server is back up.