time to find ntp server - NTP

This is a discussion on time to find ntp server - NTP ; Hi all, I have confiured an external server and the local clock. After starting ntp the ntptrace command shows 2 entries. I stop the ntp server on the external server. After 5 minuter ntptrace shows but one entry (localhost). I ...

+ Reply to Thread
Results 1 to 7 of 7

Thread: time to find ntp server

  1. time to find ntp server

    Hi all,

    I have confiured an external server and the local clock.
    After starting ntp the ntptrace command shows 2 entries.
    I stop the ntp server on the external server. After 5 minuter
    ntptrace shows but one entry (localhost).
    I restart ntp on the external server. How long should it take
    before the client becomes aware of this external server?
    System is solaris 10 galaxy with suncluster.

    # ntptrace (on cluster node 2)
    localhost: stratum 5, offset 0.000022, synch distance 0.38689
    clusternode1-priv: stratum 4, offset 0.000027, synch distance 0.01012

    after 5minutes:

    # ntptrace
    localhost: stratum 4, offset 0.000019, synch distance 0.01039

    restart xntpd on cluster node 1:

    ntptrace on clusternode 2 doesn't show clusternode1-priv anymore
    (after 30 mintes).

    Any help appreciated





  2. Re: time to find ntp server

    adirtymindisajoyforever wrote:
    > Hi all,
    >
    > I have confiured an external server and the local clock.
    > After starting ntp the ntptrace command shows 2 entries.
    > I stop the ntp server on the external server. After 5 minuter
    > ntptrace shows but one entry (localhost).
    > I restart ntp on the external server. How long should it take
    > before the client becomes aware of this external server?
    > System is solaris 10 galaxy with suncluster.
    >
    > # ntptrace (on cluster node 2)
    > localhost: stratum 5, offset 0.000022, synch distance 0.38689
    > clusternode1-priv: stratum 4, offset 0.000027, synch distance 0.01012
    >
    > after 5minutes:
    >
    > # ntptrace
    > localhost: stratum 4, offset 0.000019, synch distance 0.01039
    >
    > restart xntpd on cluster node 1:
    >
    > ntptrace on clusternode 2 doesn't show clusternode1-priv anymore
    > (after 30 mintes).
    >
    > Any help appreciated
    >


    ntptrace tells you the chain of servers the node is synchronized to,
    not the list of servers it is polling. You have apparently configured your
    local clock at the same stratum (4) as your external server,
    so cluster node 2 remained synchronized to its local clock
    after NTP was restarted on clusternode1-priv.

    Use "ntpq -p" rather than ntptrace to see the status of all your
    servers/peers.

    -Tom

  3. Re: time to find ntp server

    On 15 feb, 15:19, Tom Smith wrote:
    > adirtymindisajoyforever wrote:
    > > Hi all,

    >
    > > I have confiured an external server and the local clock.
    > > After starting ntp the ntptrace command shows 2 entries.
    > > I stop the ntp server on the external server. After 5 minuter
    > > ntptrace shows but one entry (localhost).
    > > I restart ntp on the external server. How long should it take
    > > before the client becomes aware of this external server?
    > > System is solaris 10 galaxy with suncluster.

    >
    > > # ntptrace (on cluster node 2)
    > > localhost: stratum 5, offset 0.000022, synch distance 0.38689
    > > clusternode1-priv: stratum 4, offset 0.000027, synch distance 0.01012

    >
    > > after 5minutes:

    >
    > > # ntptrace
    > > localhost: stratum 4, offset 0.000019, synch distance 0.01039

    >
    > > restart xntpd on cluster node 1:

    >
    > > ntptrace on clusternode 2 doesn't show clusternode1-priv anymore
    > > (after 30 mintes).

    >
    > > Any help appreciated

    >
    > ntptrace tells you the chain of servers the node is synchronized to,
    > not the list of servers it is polling. You have apparently configured your
    > local clock at the same stratum (4) as your external server,
    > so cluster node 2 remained synchronized to its local clock
    > after NTP was restarted on clusternode1-priv.
    >
    > Use "ntpq -p" rather than ntptrace to see the status of all your
    > servers/peers.
    >
    > -Tom


    hi,

    thanks for the reply, but before the peer disappears localhost and
    the peer node have different
    stratum (node 1 has 4, and localhost has 5).


  4. Re: time to find ntp server

    adirtymindisajoyforever wrote:
    > On 15 feb, 15:19, Tom Smith wrote:
    >> adirtymindisajoyforever wrote:
    >>> Hi all,
    >>> I have confiured an external server and the local clock.
    >>> After starting ntp the ntptrace command shows 2 entries.
    >>> I stop the ntp server on the external server. After 5 minuter
    >>> ntptrace shows but one entry (localhost).
    >>> I restart ntp on the external server. How long should it take
    >>> before the client becomes aware of this external server?
    >>> System is solaris 10 galaxy with suncluster.
    >>> # ntptrace (on cluster node 2)
    >>> localhost: stratum 5, offset 0.000022, synch distance 0.38689
    >>> clusternode1-priv: stratum 4, offset 0.000027, synch distance 0.01012
    >>> after 5minutes:
    >>> # ntptrace
    >>> localhost: stratum 4, offset 0.000019, synch distance 0.01039
    >>> restart xntpd on cluster node 1:
    >>> ntptrace on clusternode 2 doesn't show clusternode1-priv anymore
    >>> (after 30 mintes).
    >>> Any help appreciated

    >> ntptrace tells you the chain of servers the node is synchronized to,
    >> not the list of servers it is polling. You have apparently configured your
    >> local clock at the same stratum (4) as your external server,
    >> so cluster node 2 remained synchronized to its local clock
    >> after NTP was restarted on clusternode1-priv.
    >>
    >> Use "ntpq -p" rather than ntptrace to see the status of all your
    >> servers/peers.
    >>
    >> -Tom

    >
    > hi,
    >
    > thanks for the reply, but before the peer disappears localhost and
    > the peer node have different
    > stratum (node 1 has 4, and localhost has 5).
    >


    Yes, localhost has 5 because it has synced to clusternode1-priv,
    which is operating at stratum 4. After you shut that off, localhost
    goes to 4 because it is synced to its own local clock,
    which has apparently been configured with "fudge 127.127.1.? stratum 3".
    After you bring clusternode1-priv back up, it has its choice
    of its local clock at stratum 3 or clusternode1-priv at stratum 4.
    They are close, so it stays with its local clock.

    Look at "ntpq -p" and ntp.conf, and then change your ntp.conf so it has
    "fudge 127.127.1.? stratum 10" instead of stratum 3.

    -Tom


  5. Re: time to find ntp server

    >>> In article , adirtymindisajoyforever writes:

    adirtymindisajoyforever> Hi all, I have confiured an external server and the
    adirtymindisajoyforever> local clock.

    Is this machines the fallback ntp server for other machines on your network?

    --
    Harlan Stenn
    http://ntpforum.isc.org - be a member!

  6. Re: time to find ntp server

    adirtymindisajoyforever wrote:
    >
    > I have confiured an external server and the local clock.


    Why have you configured the local clock (it should never be used on leaf
    nodes, and its use on hub nodes involves tradeoffs, so it should not be
    used unless you understand them).

    >
    > restart xntpd on cluster node 1:


    xntpd is obsolete or misnamed. In this case I think it is obsolete, so
    your best chance of getting reliable information is from Sun.

  7. Re: time to find ntp server

    adirtymindisajoyforever wrote:
    > Hi all,
    >
    > I have confiured an external server and the local clock.
    > After starting ntp the ntptrace command shows 2 entries.
    > I stop the ntp server on the external server. After 5 minuter
    > ntptrace shows but one entry (localhost).
    > I restart ntp on the external server. How long should it take
    > before the client becomes aware of this external server?
    > System is solaris 10 galaxy with suncluster.
    >
    > # ntptrace (on cluster node 2)
    > localhost: stratum 5, offset 0.000022, synch distance 0.38689
    > clusternode1-priv: stratum 4, offset 0.000027, synch distance 0.01012
    >
    > after 5minutes:
    >
    > # ntptrace
    > localhost: stratum 4, offset 0.000019, synch distance 0.01039
    >
    > restart xntpd on cluster node 1:
    >
    > ntptrace on clusternode 2 doesn't show clusternode1-priv anymore
    > (after 30 mintes).
    >
    > Any help appreciated
    >
    >
    >
    >


    I switched to using NTP (available from
    http://www.ntp.org/downloads.html ) which I compiled as 64bit SPARC and
    AMD64 binaries and use with local time radio receivers. I stopped using
    the Xnptd shipped with Solaris over 6 years ago.

+ Reply to Thread