Prev: download swf video
Next: dvdbackup failure
From: balgach on 16 Jun 2005 15:07
Hello all, i am having a hard time getting ntp up and running. Right
now i am trying to just set up the server machine (which will get its
time from teh internet and then server that time to nodes in my
cluster) the /etc/ntp.conf file looks like:

restrict 127.0.0.1
restrict 128.180.38.0 mask 255.255.255.0 nomodify notrap
server -4 clock.redhat.com
server -4 us.pool.ntp.org
server -4 louie.udel.edu
server 127.127.1.0 # local clock
fudge 127.127.1.0 stratum 3 refid NIST
driftfile /etc/ntp/drift
keys /etc/ntp/keys

the server is sitting on ip 180.128.38.200 and running fedora core 2

now when i have ntpd stopped and issue something like:

[[root(a)grid00 etc]# ntpq -p louie.udel.edu
remote refid st t when poll reach delay offset
jitter
==============================================================================
-huey.udel.edu 128.175.60.175 2 u 83 1024 377 0.784 -0.512
0.733
-128.4.40.20 .GPS1. 1 u 11 64 377 0.680 -29.389
0.638
-128.4.40.19 .GPS2. 1 u 6 64 377 0.750 -29.523
0.585
*rackety.udel.ed .PPS. 1 u 231 1024 377 0.905 0.397
0.279
+ntp1.nss.udel.e .GPS. 1 u 684 1024 377 2.477 0.498
0.109
+NAVOBS1.MIT.EDU .PSC. 1 u 332 1024 377 10.441 0.780
0.202
-otc1.psu.edu .WWV. 1 u 327 1024 377 21.123 -21.654
1.752
-ntp0.usno.navy. .USNO. 1 u 221 1024 237 53.363 21.150
0.756

and it correctly sets my clock....

however when i do a service start i get:

[root(a)grid00 etc]# service ntpd start
ntpd: Synchronizing with time server: [FAILED]
Starting ntpd: [ OK ]

now in /var/log/messages i have the following lines regarding this
startup...

Jun 16 13:50:24 grid00 ntpdate[10622]: no server suitable for
synchronization found
Jun 16 13:50:24 grid00 ntpd: failed
Jun 16 13:50:24 grid00 ntpd[10627]: ntpd 4...(a)1.1161-r Thu Mar 11
11:46:39 EST 2004 (1)
Jun 16 13:50:24 grid00 ntpd[10627]: precision = 1.000 usec
Jun 16 13:50:24 grid00 ntpd[10627]: kernel time sync status 0040
Jun 16 13:50:24 grid00 ntpd: ntpd startup succeeded
Jun 16 13:50:24 grid00 ntpd[10627]: frequency initialized 0.000 PPM
from /etc/ntp/drift

i am not sure why these arent suitable for sync, since it can sync with
them when i try and do it manually. or am i just making some weird
mistake? i just want it synced up with an EST server. thanks.

Cheers,

Adam.

From: Douglas Mayne on 16 Jun 2005 16:10
On Thu, 16 Jun 2005 12:07:59 -0700, balgach wrote:

> Hello all, i am having a hard time getting ntp up and running. Right
> now i am trying to just set up the server machine (which will get its
> time from teh internet and then server that time to nodes in my
> cluster) the /etc/ntp.conf file looks like:
>
> restrict 127.0.0.1
> restrict 128.180.38.0 mask 255.255.255.0 nomodify notrap
dislexic number order in dotted quad?

> server -4 clock.redhat.com
> server -4 us.pool.ntp.org
> server -4 louie.udel.edu
> server 127.127.1.0 # local clock
> fudge 127.127.1.0 stratum 3 refid NIST
> driftfile /etc/ntp/drift
> keys /etc/ntp/keys
>
> the server is sitting on ip 180.128.38.200 and running fedora core 2
dislexic number order in dotted quad?

>
> now when i have ntpd stopped and issue something like:
>
> [[root(a)grid00 etc]# ntpq -p louie.udel.edu
> remote refid st t when poll reach delay offset
> jitter
> ==============================================================================
> -huey.udel.edu 128.175.60.175 2 u 83 1024 377 0.784 -0.512
> 0.733
> -128.4.40.20 .GPS1. 1 u 11 64 377 0.680 -29.389
> 0.638
> -128.4.40.19 .GPS2. 1 u 6 64 377 0.750 -29.523
> 0.585
> *rackety.udel.ed .PPS. 1 u 231 1024 377 0.905 0.397
> 0.279
> +ntp1.nss.udel.e .GPS. 1 u 684 1024 377 2.477 0.498
> 0.109
> +NAVOBS1.MIT.EDU .PSC. 1 u 332 1024 377 10.441 0.780
> 0.202
> -otc1.psu.edu .WWV. 1 u 327 1024 377 21.123 -21.654
> 1.752
> -ntp0.usno.navy. .USNO. 1 u 221 1024 237 53.363 21.150
> 0.756
>
> and it correctly sets my clock....
>
> <snip>
Check numbers that you have entered in the config file (inline comments
above.)

ntpdate may need to set the clock before the ntpd can start. This is
because your local clock is too far off to be acceptable to ntp.

Check the documentation online at:
http://www.eecis.udel.edu/~mills/ntp/html/debug.html

From: Steffen Kluge on 17 Jun 2005 10:36
balgach(a)gmail.com wrote:
> server 127.127.1.0 # local clock
> fudge 127.127.1.0 stratum 3 refid NIST

Are you using a local reference clock?? If not, drop that. It's not
going to help.

> [root(a)grid00 etc]# service ntpd start
> ntpd: Synchronizing with time server: [FAILED]
> Starting ntpd: [ OK ]
>
> now in /var/log/messages i have the following lines regarding this
> startup...
>
> Jun 16 13:50:24 grid00 ntpdate[10622]: no server suitable for
> synchronization found

The ntpd startup script runs ntpdate to sync off servers listed in
/etc/ntp/step-tickers (as you can easily verify yourself - startup
scripts are text files :)

What do you have in step-tickers?

Cheers
Steffen.
From: Christopher Browne on 17 Jun 2005 12:41
A long time ago, in a galaxy far, far away, Steffen Kluge <kluge(a)dotnet.org> wrote:
> balgach(a)gmail.com wrote:
>> server 127.127.1.0 # local clock
>> fudge 127.127.1.0 stratum 3 refid NIST
>
> Are you using a local reference clock?? If not, drop that. It's not
> going to help.

I generally set up the local clock, but put it to a *WAY* lower
stratum, usually stratum 10. That way, *anything* that I can connect
to will be considered superior to the local clock.

Happily, that ultimately feeds back to set the local clock, so that if
I have some temporary network outage, the local clock *can* do some
good, and let NTP keep running during the outage...

>> [root(a)grid00 etc]# service ntpd start
>> ntpd: Synchronizing with time server: [FAILED]
>> Starting ntpd: [ OK ]
>> now in /var/log/messages i have the following lines regarding this
>> startup...
>> Jun 16 13:50:24 grid00 ntpdate[10622]: no server suitable for
>> synchronization found
>
> The ntpd startup script runs ntpdate to sync off servers listed in
> /etc/ntp/step-tickers (as you can easily verify yourself - startup
> scripts are text files :)
>
> What do you have in step-tickers?

Yeah, you'll need to have some kind of list of servers.

- It is usually a good idea to put pool.ntp.org on the list. That
does a round robin access to a pool of NTP servers.

- You could do a traceroute to some place far away, and grab the first
few hosts you hit, likely within your ISP's infrastructure, and
try them; it is common for Cisco routers to run NTP, and they will
Definitely will be "nearby on the network."

- There may be some favored local servers you know of, either inside
local network or outside...
--
output = ("cbbrowne" "@" "gmail.com")
http://linuxdatabases.info/info/internet.html
"Many companies that have made themselves dependent on [the equipment
of a certain major manufacturer] (and in doing so have sold their soul
to the devil) will collapse under the sheer weight of the unmastered
complexity of their data processing systems."
-- Edsger W. Dijkstra, SIGPLAN Notices, Volume 17, Number 5
 | 
Pages: 1
Prev: download swf video
Next: dvdbackup failure