Whenever the server is rebooted, I’ve noticed that the time goes a bit screwy and jumps forward an hour, having tried a few different settings I found that it wasn’t just this server but others were having the same problem.
Apparently on some Dell servers running Linux, the command ‘/sbin/hwclock –systohc’ (to write the current system time into the hardware clock) fails with the error:
select() to /dev/rtc to wait for clock tick timed out
The solution is to add –directisa to the command.
/sbin/hwclock –systohc –directisa
If you’re running Fedora Core and possibly RedHat, you can add this to a configuration file so that it works when shutting down or rebooting the server. Just add the line:
CLOCKFLAGS=”–directisa”
at the bottom of /etc/sysconfig/clock (or edit the existing CLOCKFLAGS line if there is one).
Excellent! Thanks for that. I have a few dells at work that have always given me this problem and so far I have always adjusted the clock manually on reboot (which isn’t that often, but still annoying)
Hmm. Since hwclock will default to trying –directisa anyway if it can’t read /dev/rtc then perhaps it should also default to trying again with –directisa if it gets a clock tick timeout?
Where do I put this line for a Debian system??
CLOCKFLAGS=”–directisa”
/etc/sysconfig/clock doesn’t exist on Debian. What’s the equivalent file?