f



ntpq -p show refid as .INIT. even my NTP Servers are synchronized properly.

hi,

We are configured NTP in Multiple servers and all the servers are
getting snchronized properly.We are having primary and Secondary NTP
servers, where as primary NTP server is refering its LOCAL BIOS as
reference clock.

And Secondary NTP server refering Primary NTP server as reference
clock. And other multiple NTP servers refering Primary as reference
clock and if primary goes down they will be pointing secondary as
reference clock automatically.

Problem:

When i execute ntpq -p the output shows refid as .INIT. ?
How to resolve this?
Why it is showing?
Tell me how much time it will take to synchronize to the secondary
server when primary goes down?
0
Andhu
10/4/2009 6:02:02 PM
comp.protocols.time.ntp 4895 articles. 2 followers. Post Follow

5 Replies
3187 Views

Similar Articles

[PageSpeed] 50

Andhu <aravind.arjunan@gmail.com> writes:

>hi,

>We are configured NTP in Multiple servers and all the servers are
>getting snchronized properly.We are having primary and Secondary NTP
>servers, where as primary NTP server is refering its LOCAL BIOS as
>reference clock.

What is "local Bios"? do you mean the Real Time Clock (the little clock
chip on the motherboard)? That is likely to
out by many seconds between the various machines. Ie, ntp will find that
there there are no "true tickers" Just a random assortment of bad
clocks. 

>And Secondary NTP server refering Primary NTP server as reference
>clock. And other multiple NTP servers refering Primary as reference
>clock and if primary goes down they will be pointing secondary as
>reference clock automatically.

>Problem:

>When i execute ntpq -p the output shows refid as .INIT. ?
>How to resolve this?
>Why it is showing?
>Tell me how much time it will take to synchronize to the secondary
>server when primary goes down?
0
Unruh
10/4/2009 6:42:58 PM
On Oct 4, 11:42=A0pm, Unruh <unruh-s...@physics.ubc.ca> wrote:
> Andhu <aravind.arju...@gmail.com> writes:
> >hi,
> >We are configured NTP in Multiple servers and all the servers are
> >getting snchronized properly.We are having primary and Secondary NTP
> >servers, where as primary NTP server is refering its LOCAL BIOS as
> >reference clock.
>
> What is "local Bios"? do you mean the Real Time Clock (the little clock
> chip on the motherboard)? That is likely to
> out by many seconds between the various machines. Ie, ntp will find that
> there there are no "true tickers" Just a random assortment of bad
> clocks.
>
>
>
> >And Secondary NTP server refering Primary NTP server as reference
> >clock. And other multiple NTP servers refering Primary as reference
> >clock and if primary goes down they will be pointing secondary as
> >reference clock automatically.
> >Problem:
> >When i execute ntpq -p the output shows refid as .INIT. ?
> >How to resolve this?
> >Why it is showing?
> >Tell me how much time it will take to synchronize to the secondary
> >server when primary goes down?- Hide quoted text -
>
> - Show quoted text -

hi Unruh

Thanks for your reply.
Local BIOS is nothing but the Hardware clock , yes it is the Real time
clock which comes in Motherboard.

What is true tickers?

You are saying that it is not recommend to keep the source reference
clock as Real time clock?

with regards
aravind
0
Andhu
10/4/2009 7:01:30 PM
On 2009-10-04, Andhu <aravind.arjunan@gmail.com> wrote:

[snip: time island with primary and secondary time servers]

> Problem:
>
> When i execute ntpq -p the output shows refid as .INIT. ?

Where are you running this command?

It may be that you are not allowing enough time to elapse after ntpd is
started.

For ntp-stable using ntpd's defaults a server which is using the
Undisciplined Local Clock (127.127.1.x) as its time "source" will not be
able to serve time to others until it has been running for approximately
200 seconds (3 1/3 minutes). This time may be reduced to approximately
50 seconds using the minpoll command.

> How to resolve this?

Fix your configuration and/or allow ntpd to run for a long enough time
before using 'ntpq -p'.

> Why it is showing?

We won't be able to tell you that until we see a sample configuration
for one of your servers and one of your clients.

> Tell me how much time it will take to synchronize to the secondary
> server when primary goes down?

The "client ntpd" will keep the primary server as its "sys_peer" (in
other words, will continue to be synced to the primary server) until
that primary server becomes unreachable.

It takes a maximum of 8 polls for a time source to become unreachable.

For unicast (or "client/server") associations, depending on the poll
interval being used by the client ntpd it will take somewhere between
8.5 minutes  and 2.27 hours for a remote time server to become
unreachable.

-- 
Steve Kostecke <kostecke@ntp.org>
NTP Public Services Project - http://support.ntp.org/
0
Steve
10/4/2009 7:19:19 PM
On Oct 5, 12:19=A0am, Steve Kostecke <koste...@ntp.org> wrote:
> On 2009-10-04, Andhu <aravind.arju...@gmail.com> wrote:
>
> [snip: time island with primary and secondary time servers]
>
> > Problem:
>
> > When i execute ntpq -p the output shows refid as .INIT. ?
>
> Where are you running this command?
>
> It may be that you are not allowing enough time to elapse after ntpd is
> started.
>
> For ntp-stable using ntpd's defaults a server which is using the
> Undisciplined Local Clock (127.127.1.x) as its time "source" will not be
> able to serve time to others until it has been running for approximately
> 200 seconds (3 1/3 minutes). This time may be reduced to approximately
> 50 seconds using the minpoll command.
>
> > How to resolve this?
>
> Fix your configuration and/or allow ntpd to run for a long enough time
> before using 'ntpq -p'.
>
> > Why it is showing?
>
> We won't be able to tell you that until we see a sample configuration
> for one of your servers and one of your clients.
>
> > Tell me how much time it will take to synchronize to the secondary
> > server when primary goes down?
>
> The "client ntpd" will keep the primary server as its "sys_peer" (in
> other words, will continue to be synced to the primary server) until
> that primary server becomes unreachable.
>
> It takes a maximum of 8 polls for a time source to become unreachable.
>
> For unicast (or "client/server") associations, depending on the poll
> interval being used by the client ntpd it will take somewhere between
> 8.5 minutes =A0and 2.27 hours for a remote time server to become
> unreachable.
>
> --
> Steve Kostecke <koste...@ntp.org>
> NTP Public Services Project -http://support.ntp.org/

hi,

When i execute ntpq -p the output shows refid as .INIT. ?

Where are you running this command?

I am  executing this command in NTP Server and in my NTP Clients.


It may be that you are not allowing enough time to elapse after ntpd
is
started.

So you mean to say that we need to execute the command ntpq -p after
the NTP server get stable?

As i mentioned it was in production and these NTP Servers were running
for very long time. And all my NTP Servers
are getting synchronized from the Primary NTP Server.

I found in some documents that this .INIT. is the kiss code error and
it means that my associaition server is not synchronized for the first
time.

For ntp-stable using ntpd's defaults a server which is using the
Undisciplined Local Clock (127.127.1.x) as its time "source" will not
be
able to serve time to others until it has been running for
approximately
200 seconds (3 1/3 minutes). This time may be reduced to
approximately
50 seconds using the minpoll command.

From the above you mean to say if the source is the Undisciplined
Local Clock (127.127.1.x) it wont serve time until it  has been
running for 3 1/3 minutes.

But my NTP Server is running for very long time. But still am getting
the refid as .INIT. when i execute ntpq -p command.

And regarding the time interval of ntpd client from changing the
synchronization from Primary to Secondary when the primary becaome
unreachable . As per you it will take maximum of 8 polls but it is
taking more than 1 hours to change it secondary NTP server if the
primary goes down.

with regards
A.Aravind

0
Andhu
10/5/2009 6:34:45 AM
Andhu wrote:
> On Oct 5, 12:19 am, Steve Kostecke <koste...@ntp.org> wrote:
>> On 2009-10-04, Andhu <aravind.arju...@gmail.com> wrote:
>>
>> [snip: time island with primary and secondary time servers]
>>
>>> Problem:
>>> When i execute ntpq -p the output shows refid as .INIT. ?
>> Where are you running this command?
>>
>> It may be that you are not allowing enough time to elapse after ntpd is
>> started.
>>
>> For ntp-stable using ntpd's defaults a server which is using the
>> Undisciplined Local Clock (127.127.1.x) as its time "source" will not be
>> able to serve time to others until it has been running for approximately
>> 200 seconds (3 1/3 minutes). This time may be reduced to approximately
>> 50 seconds using the minpoll command.
>>
>>> How to resolve this?
>> Fix your configuration and/or allow ntpd to run for a long enough time
>> before using 'ntpq -p'.
>>
>>> Why it is showing?
>> We won't be able to tell you that until we see a sample configuration
>> for one of your servers and one of your clients.
>>
>>> Tell me how much time it will take to synchronize to the secondary
>>> server when primary goes down?
>> The "client ntpd" will keep the primary server as its "sys_peer" (in
>> other words, will continue to be synced to the primary server) until
>> that primary server becomes unreachable.
>>
>> It takes a maximum of 8 polls for a time source to become unreachable.
>>
>> For unicast (or "client/server") associations, depending on the poll
>> interval being used by the client ntpd it will take somewhere between
>> 8.5 minutes  and 2.27 hours for a remote time server to become
>> unreachable.
>>
>> --
>> Steve Kostecke <koste...@ntp.org>
>> NTP Public Services Project -http://support.ntp.org/
> 
> hi,
> 
> When i execute ntpq -p the output shows refid as .INIT. ?
> 
> Where are you running this command?
> 
> I am  executing this command in NTP Server and in my NTP Clients.
> 
> 
> It may be that you are not allowing enough time to elapse after ntpd
> is
> started.
> 
> So you mean to say that we need to execute the command ntpq -p after
> the NTP server get stable?
> 
> As i mentioned it was in production and these NTP Servers were running
> for very long time. And all my NTP Servers
> are getting synchronized from the Primary NTP Server.
> 
> I found in some documents that this .INIT. is the kiss code error and
> it means that my associaition server is not synchronized for the first
> time.
> 
> For ntp-stable using ntpd's defaults a server which is using the
> Undisciplined Local Clock (127.127.1.x) as its time "source" will not
> be
> able to serve time to others until it has been running for
> approximately
> 200 seconds (3 1/3 minutes). This time may be reduced to
> approximately
> 50 seconds using the minpoll command.
> 
> From the above you mean to say if the source is the Undisciplined
> Local Clock (127.127.1.x) it wont serve time until it  has been
> running for 3 1/3 minutes.
> 
> But my NTP Server is running for very long time. But still am getting
> the refid as .INIT. when i execute ntpq -p command.
> 
> And regarding the time interval of ntpd client from changing the
> synchronization from Primary to Secondary when the primary becaome
> unreachable . As per you it will take maximum of 8 polls but it is
> taking more than 1 hours to change it secondary NTP server if the
> primary goes down.
> 
> with regards
> A.Aravind
> 

As he said, please give cofiguration details, ntp.conf examples
from primary server, secondary server and client.

 From logs last month where I struggled to find a 4 day period
when I'd not lost internet connection there was at least one
outage of 6 hrs but ntpd hardly noticed (polls on servers are
up at 68m and 137m for selected sources).

servers here have
	server 127.127.1.0
	fudge 127.127.1.0 stratum 12

David
0
David
10/5/2009 9:54:48 AM
Reply: