Skip to main content

Updating and port 37

Comments

20 comments

  • quietFinn

    What happens if you do this in command line:
    telnet rdate.cpanel.net 37

    It should be like this:
    [root@srv ~]# telnet rdate.cpanel.net 37
    Trying 208.74.121.43...
    Connected to rdate.cpanel.net.
    Escape character is '^]'.
    ▒▒▒PConnection closed by foreign host.


    0
  • Simon Blackburn

    I get

    Bash: telnet: command not found

    0
  • Simon Blackburn

    I don’t think rdate is installed. Another check returned that Chrony is installed but how do I get it to use Chrony instead of rdate.

    0
  • quietFinn

    Install telnet:
    yum install telnet

    and try again.

    The error message you get suggests that your server can't connect to rdate.cpanel.net

     

    0
  • Simon Blackburn

    telnet rdate.cpanel.net 37
    Trying 208.74.123.15.
    -elnet: connect to address 208.74.123.15: Connection ret
    Trying 208.74.121.36.
    telnet: connect to address 208.74.121.36: Connection ref
    Trying
    208.74.123.23...
    telnet: connect to address 208.74.123.23: Connection ref
    Trying 208.74.121.43..
    telnet: connect to address 208.74.121.43: Connection ref Used

    0
  • Simon Blackburn

    That should say refused not ref used.

    0
  • quietFinn

    https://support.cpanel.net/hc/en-us/articles/4414037344535-rdate-reports-an-error-when-upcp-runs

    "The rdate script adjusts the system clock on servers not running an NTP daemon. The errors occur when the server is unable to contact rdate.cpanel.net on TCP port 37. This can occur if the server is unable to resolve rdate.cpanel.net or if outbound TCP port 37 is blocked."

     

    0
  • Simon Blackburn

    I know port 37 isn’t blocked.

    Do I need rdate if I have Chrony installed and running?

    0
  • quietFinn

    I closed port 37 and this is what I get:

    [root@vps1 log]# telnet rdate.cpanel.net 37
    Trying 208.74.123.15...
    telnet: connect to address 208.74.123.15: Connection refused
    Trying 208.74.121.36...
    telnet: connect to address 208.74.121.36: Connection refused
    Trying 208.74.123.23...
    telnet: connect to address 208.74.123.23: Connection refused
    Trying 208.74.121.43...
    telnet: connect to address 208.74.121.43: Connection refused

     

    0
  • Simon Blackburn

    I did a check earlier and it said it was open. I’ll reinstall CSF and see what it says then. Thanks.

    0
  • Simon Blackburn

    I've installed CSF again and configured it. Port 37 is open for TCP_OUT. 

    I've ran telnet rdate.cpanel.net 37 again and it still cannot connect.

    0
  • cPRex Jurassic Moderator

    It's also possible the port is blocked outside of your server so if you confirm things are working will in CSF you may need to speak with your host.

    0
  • dexus

    We actually don't need this rdate script. This archaic script is not updated for EL7+ servers, so it is not checking for chrony service, only ntp, which is replaced with chrony a long time ago, and this script is just making problems because of harsh time changes on all servers every day...

    systemd-journald[312416]: Time jumped backwards, rotating.

    For example, we are losing journal logs al most every day, on all servers, because of this script.

    @cPanel Please open a bug for this and fix this script to not update time if chrony is running.

     

    0
  • dexus

    As a temporary solution we can just do this on all cPanel servers that have chronyd service running...

    mv /usr/local/cpanel/scripts/rdate /usr/local/cpanel/scripts/rdate.bak
    ln -s /bin/true /usr/local/cpanel/scripts/rdate

     

    1
  • cPRex Jurassic Moderator

    dexus - it sounds like there is something else happening on your machine that is causing the time to drift, as I don't have see any other issues or complaints with the rdate script at this time.  Would you be able to submit a ticket so this can be investigated directly on the server?

    1
  • Simon Blackburn

    I took a snapshot of my server and reinstalled everything from scratch. Thanks for everyone’s help. Don’t know why the original installation was corrupted, it was all done via SSH.

    0
  • cPRex Jurassic Moderator

    Simon Blackburn- glad to hear things are working well now!

    1
  • dexus

    cPRex We have the same issue on all servers, at exact time when rdate is run, and also after "replacing" rdate script, as I suggested, there is no such issue, so it is 100% the cause for the issue.

    Anyway that script is made for use on CentOS 6 and older OS versions, and it must be updated to also check if chronyd is running, like it does for ntpd, and exit if ntpd is running. It should do the same for chrony.

    It does not make sense to use rdate to update time on servers that have chronyd running.

     

    0
  • cPRex Jurassic Moderator

    dexus - can you submit at ticket about this?  I don't have any other similar reports of this causing an issue so we'd likely need to see it in action on one of your affected systems before we could say for sure what is happening.

    0
  • dexus

    Sorry, I don't have time for that, and there is really nothing to be checked on server. Such time related issues are expected when using rdate on newer OS. That is just how rdate works, because it was made for OS before journalctl and chrony and it is unfortunately still in production.

    That script is for CentOS 6 and older OS versions, and it is not updated to check for chrony.

    Just discuss this with any of your senior system engineers or developers.

     

    0

Please sign in to leave a comment.