Network time protocol (NTP) client configuration help needed urgently

Network time protocol (NTP) client configuration help needed urgently

Post by sridh » Sat, 28 Jun 2003 15:14:40



Hi,

I want to configure my redhat linux 8 machines for Network time
protocol(NTP) clients and my NTP server will be in solaris 8
machine.How to configure NTP client in redhat linux 8.Please help me
in this.I am in need of urgent.

Thanks in advance

Regards

 
 
 

Network time protocol (NTP) client configuration help needed urgently

Post by ujudzm.. » Sat, 28 Jun 2003 15:54:22


|I want to configure my redhat linux 8 machines for Network time
|protocol(NTP) clients and my NTP server will be in solaris 8
|machine.How to configure NTP client in redhat linux 8.Please help me
|in this.I am in need of urgent.

Have you tried a web search for a NTP howto? Plus there is a lot of
documentation at the udel NTP site.

 
 
 

Network time protocol (NTP) client configuration help needed urgently

Post by Seth H Holme » Sat, 28 Jun 2003 21:56:31



> Hi,

> I want to configure my redhat linux 8 machines for Network time
> protocol(NTP) clients and my NTP server will be in solaris 8
> machine.How to configure NTP client in redhat linux 8.Please help me
> in this.I am in need of urgent.

man ntpd

Should have everything you need to know.

--
Seth H Holmes

 
 
 

1. Problem with ntp on solaris (also posted on comp.protocols.time.ntp)

Hello,

I am trying to setup ntp using a hopf 6021 radio clock. We have the
following configuration:

NTP V4.1.71 on Sunos V5.8/solaris v8 compiled with gcc 2.95.3. Because
our hopf clock is strapped to generate a y2k compliant string i had to
make some minor changes in the file libparse/clk_hopf6021.c (two
additional bytes for the year string).

After starting the daemon it looks like it works all right, the stratum
becomes 1 after about 8 minutes. But after about 15 minutes the syslog
is flooded with messages like:

Oct 31 11:34:39 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 2!
Oct 31 11:34:39 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 3!
Oct 31 11:34:39 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 4!
Oct 31 11:34:39 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 5!
Oct 31 11:34:39 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 4!
Oct 31 11:34:40 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 2!
Oct 31 11:34:40 xxxxxx ntpd[18241]: [ID 702911 daemon.error]
input_handler: handler_count is 3!

I also see sometimes the messages:

Oct 31 11:34:40 xxxxxx ntpd[18241]: [ID 702911 daemon.error] clock read
fd 8: Resource temporarily unavailable
Oct 31 11:34:43 xxxxxx ntpd[18241]: [ID 702911 daemon.error] select(9,
#-###----, 0L, 0L, &0.000000) error: Interrupted system call

These messages continue for several minutes and sometimes for over an
hour. Then it becomes quiet again for some time.

And one more (and probably more serious) problem is that after several
hours of running i get the following messages:

Oct 31 17:55:17 xxxxxx su: [ID 643653 kern.notice] NOTICE: su0: ring
buffer overflow
Oct 31 17:55:17 xxxxxx
Oct 31 17:55:17 xxxxxx su: [ID 643653 kern.notice] NOTICE: su0: ring
buffer overflow
Oct 31 17:55:17 xxxxxx
Oct 31 17:55:17 xxxxxx su: [ID 643653 kern.notice] NOTICE: su0: ring
buffer overflow
Oct 31 17:55:17 xxxxxx

About 120 MB of these overnight! As far as i know the su0 device
corresponds to the serial port where the clock is connected to.

At this moment i checked NTP and it showed the following:

ntpq> clocklist
status=0001 clk_okay, last_clk_noreply
device="HOPF 6021", timecode="\x02CB16004431102001\x0a\x0d\x03",
poll=922, noreply=455, badformat=2, baddata=0, fudgetime1=0.000,
fudgetime2=0.000, stratum=0, refid=DCF, flags=0,
refclock_time="bf8a9eac.00000000  Wed, Oct 31 2001 16:00:44.000 UTC",
refclock_status="UTC DISPLAY; TIME CODE",
refclock_format="hopf Funkuhr 6021",
refclock_states="*NOMINAL: 00:34:22 (3.47%); NO RESPONSE: 15:55:54
(96.52%); ILLEGAL DATE: 00:00:01 (0.00%); running time: 16:30:17"
ntpq>

Especially the high percentage of no response is probably related to the
above errors.

Has anybody seen this? My experience with Solaris is not very much but
this looks related to the way Solaris handles asynchronous IO
operations. Could this be a Solaris problem or NTP?

Thanks in advance,

Oswald

2. From address differences

3. NTP (Network Time Protocol) Questions

4. 4.6.27 libs: now login won't work!

5. NTP (Network Time Protocol)

6. Learning CVS

7. find, cflow usage problem.

8. NTP (NEtwork Time Protocol) setting up?

9. NTP: Network Time Protocol

10. NTP Network Time Protocol for Linux

11. IP Service 'ntp' (number 123 - Network Time Protocol) for SPARC Solaris 2.4?

12. How does "Time Synch Protocol" relate to "Network Time Protocol"?