ntpd on Fedora Core 6 with high negative time reset values
- by Mark White
The basic problem is we have a FC6 server instance running on a virtual machine, and the system time seems to have been slowly varying until it is now causing a problem. The server runs 24/7 and has been up for 155 days.
It has been changed to show GMT, and reports the time as (example) 00:15:15 GMT whereas the actual time is 00:00:00 GMT. This is an offset of 915 seconds.
selinux has been changed to 'setenforce 0' for testing and I am running as root.
I stop the ntpd service and change the time in System|Administration|Date & Time. The time still shows the same with 'date' in bash. There are no error logs.
I change the date with 'date --set' in bash. The response confirms the changed date. I run 'date' and the incorrect date is shown. There are no error logs.
I start the ntpd service and /var/log/messages shows success with 'time reset -915.720139s'. The date remains unchanged.
ntpq -p shows three three time servers all have offsets of around -915 seconds.
I stop ntpd service and try 'ntpd -gqx' and get the same result as above - success, but a large negative time reset.
I've tried varying combinations of the above, and a few more settings in System|Administration|Date & Time - no change. I just need to reset the system time to GMT. No offset. But I can't wait for ntpd to slew the time over the next few weeks.
Any advice is welcome, cheers! Surely this shouldn't be this difficult...
Mark...
