| 1 |
Drew Tomlinson <drew@××××××××××××××.net> writes: |
| 2 |
|
| 3 |
> The time server is a FreeBSD 6.0 box on my network. My other FreeBSD |
| 4 |
> box and two Windows boxes get time from it just fine. Even the Gentoo |
| 5 |
> box will set its clock with "ntpd -gq". I am currently using this |
| 6 |
> brute force method via a cron job as a temporary workaround. |
| 7 |
> |
| 8 |
> Any ideas on what might have caused this recent change in behavior? |
| 9 |
|
| 10 |
It might be that the clock got off by more than ntp is willing to |
| 11 |
adjust. As I recall there is a threshold above which ntp will not |
| 12 |
go. Hopefully someone more knowledgable might confirm that. |
| 13 |
|
| 14 |
To cure that sort of problem here I run ntp-client at boot. It sets |
| 15 |
the time by any amount I think. So time gets set right on boot then |
| 16 |
ntp will keep it in good order. Next boot up ntp-client comes in |
| 17 |
ahead of ntp and sets the clock before ntp gets to it, so the |
| 18 |
too-large discrepancy never occurs. |
| 19 |
|
| 20 |
I seem to recall that the too-large discrepancy is not really that |
| 21 |
large. I remember thinking it seemed kind of small to be a problem. |
| 22 |
|
| 23 |
-- |
| 24 |
gentoo-user@g.o mailing list |
Archives