clock problem: Further information

From: Guy Dallaire <dallaire_at_megatoon.com>
Date: Wed, 10 Jul 1996 11:37:47 -0400

See my previous post please.

- Both servers are running xntpd.
- Both servers were shut down at the same time ytesterday before the
problems began.
- One of the servers(hostname=dgeux1) has the following peers in
/etc/xntpd.conf file:

     127.127.1.1

  (I have NO machine with such an IP address !)

- The other server (dgeux1) has the following peers:

    dgeux2 (Self !)
    dgeux1

Looks like dgeux1's clock went "boom" and dgeux2 is synchronizing with
dgeux1's clock... Am I right ? What could have caused such a thing and how
can I prevent this ? In rc.config, there is a parameter telling xntpd to
"adjust the system time even if the drif is > 1000 seconds (-g parameter)".
Looks like it is doing it. Isn't it dangerous to set the system time in
multi-user mode (Manual says that user could experiment corruption in file
creation/modification times) ? If so, why is xntpd doing it ?

Do I have to go single user to reset the date on dgeux1 ?

Thanks
Received on Wed Jul 10 1996 - 18:54:44 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:46 NZDT