1 |
Steve Herber posted <Pine.LNX.4.64.0603140110330.17364@×××××.com>, |
2 |
excerpted below, on Tue, 14 Mar 2006 01:26:10 -0800: |
3 |
|
4 |
> I have a nice new A8N-SLI Premium system and I have two problems with it. |
5 |
> The second problem I have is that ntpd fails with the error: |
6 |
> |
7 |
> Mar 13 08:41:54 fester ntpd[16778]: ntpd 4.2.0a@1.1191-r Thu Feb 23 |
8 |
> 17:17:02 PST 2006 (1) |
9 |
> Mar 13 08:41:54 fester ntpd[16778]: precision = 1.000 usec |
10 |
> Mar 13 08:41:54 fester ntpd[16778]: Listening on interface wildcard, |
11 |
> 0.0.0.0#123Mar 13 08:41:54 fester ntpd[16778]: Listening on interface lo, |
12 |
> 127.0.0.1#123 |
13 |
> Mar 13 08:41:54 fester ntpd[16778]: Listening on interface eth1, |
14 |
> 192.168.168.6#123 |
15 |
> Mar 13 08:41:54 fester ntpd[16778]: kernel time sync status 0040 |
16 |
|
17 |
A search of the ntp documentation (in /usr/doc/ntp-<ver>/html/*) for sync |
18 |
status returns the error docs page. It refers to timex.h from the ntp |
19 |
sources tarball. You probably have that tarball in your portage distdir. |
20 |
A quick look here at the file... exerpted: |
21 |
|
22 |
/* |
23 |
* Status codes (timex.status) |
24 |
*/ |
25 |
[snip] |
26 |
#define STA_INS 0x0010 /* insert leap (rw) */ |
27 |
#define STA_DEL 0x0020 /* delete leap (rw) */ |
28 |
#define STA_UNSYNC 0x0040 /* clock unsynchronized (rw) */ |
29 |
#define STA_FREQHOLD 0x0080 /* hold frequency (rw) */ |
30 |
|
31 |
So... that would appear to simply say status unsynchronized. Not a |
32 |
problem as that's the way it's supposed to start. In fact, I figured |
33 |
that's what it meant, as mine always starts that way, but the status |
34 |
changes, later. |
35 |
|
36 |
Thus, either ntpd isn't failing after all, but still running and will |
37 |
eventually sync, or if it's failing, the log as posted doesn't reveal why. |
38 |
|
39 |
A couple suggestions. |
40 |
|
41 |
1: Check your timezone info. If you have your clock set to local time |
42 |
but the timezone is set to UTC, assuming you aren't /at/ UTC local time, |
43 |
you'll be off by several hours and ntpd will normally refuse to touch the |
44 |
clock untl you set it to something closer to what ntpd can grok. |
45 |
|
46 |
2: It's always best to run ntp-client as well as ntpd. The init ordering |
47 |
will order ntp-client before ntpd if they are both in the same initlevel, |
48 |
which is what you want. ntp-client does a quick-check and approximate |
49 |
sync, stepping the clock as necessary to get it /close/ to the right time. |
50 |
It is, however, a one-shot deal. ntpd then takes over and makes fine |
51 |
adjustments as necessary to sync the clock more closely, then tracks it to |
52 |
ensure it stays synced. By running ntp-client first, you ensure the clock |
53 |
is close enough to correct that ntpd shouldn't protest. It will still |
54 |
start as status 0040, but will sync up, usually within 15 minutes or so. |
55 |
|
56 |
-- |
57 |
Duncan - List replies preferred. No HTML msgs. |
58 |
"Every nonfree program has a lord, a master -- |
59 |
and if you use the program, he is your master." Richard Stallman in |
60 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
61 |
|
62 |
|
63 |
-- |
64 |
gentoo-amd64@g.o mailing list |