11:27 a.m.
If I block WAN connection to a server will it find a time reference on
the LAN? There's usually a workstation connected and getting it's clock
updated but it looks like the server is not? I haven't found the answer
in the man pages ... I was looking for a configuration setting perhaps.
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
3:14 a.m.
On 01/03/15 20:55, g wrote:
laughing at yourself or using syncing servers to a workstation.;-)
hi Bob,
along with what Ed Greshko suggested,,,
being that you have mentioned the internet time cost with being on
satellite, you prefer to limit running ntp for clock syncing.
so what is wrong with WWV? surely you can receive WWV on 10 meters from
where you are.
and if so, you can tweak 'drift rate' /etc/adjtime. just check WWV time
every 24 hrs at same time for a week and tweak 'drift rate' to compensate.
after 7 days, check every other day until you get drift to minimal.
lines in /etc/adjtime are;
line 1 numbers:
1st= systematic drift rate in seconds per day,
floating point decimal.
2nd= resulting number of seconds since 1969 UTC of most recent
adjustment or calibration, decimal integer.
3rd= zero, for compatibility with clock(8), decimal integer.
line 2 number:
resulting number of seconds since 1969 UTC of most recent
calibration. zero if there has been no calibration yet
or it is known that any previous calibration is moot.
ie, because the hardware clock has been found, since that
calibration, not to contain a valid time, decimal integer.
line 3:
"UTC" or "LOCAL". setting of the hardware.
you could also use 'hwclock' to adjust /etc/adjtime. see "man hwclock"
for how to use. also, a good article;
http://www.sanfoundry.com/5-hwclock-command-usage-examples-linux/
Good morning George:
Glad to hear from you, you've been conspicuous by your absence, at least
from the lists I read. I wish you a healthy, happy, and prosperous new year.
We've made it through the Christmas holiday and life should gradually
return to near normal here.
Linux offers some sophisticated schemes for keeping time and I prefer to
use them, they have been working for me for a long time although I have
been aware of a problem with the servers and I want to try and fix those
so they work the same way.
I don't think I need microsecond accuracy, just would like them all to
agree to the nearest second perhaps. I was seeing a couple of minutes
difference in the samba server the other day which made it hard to
correlate date between the server and client logs.
WWV and WWVB are available and could be used but that requires constant
attention from me and I would prefer avoiding that unless there's no
other way.
I now have most of the samba problems sorted out, I still can't run
firewalld on the server end though and haven't found why yet. Firewalld
and selinux work on everything else but samba is really a nightmare to
deal with, too many Windows terms, etc.
Well I've had my coffee and two Keebler oatmeal cookies so now I can go
back to bed until the sun comes over the edge, Addy, my weimaraner will
probably want to go out first but I can still get a few zzz's.
Later,
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
7:01 a.m.
New subject: chrony lan - OT -
Allegedly, on or about 05 January 2015, Bob Goodwin sent:
Crony/ntp appears to be giving the expected results this morning
after
being left to do whatever it does overnight and rebooting the client
again. So I conclude that although I can do systemctl restart chrony
that doesn't initiate an update of the system time correction? Not
knowing how to do that makes it difficult to test ...
I'm not really familiar with chrony, but did get NTP sailing smoothly in
the past. I would expect it to work in a similar way to how NTP was
configured, and there were three common ways systems were set up to use
NTP:
1. Start up the NTP daemon, and if the clock wasn't too far off wrong,
correct it, and keep it running on time. i.e. It's not a set once and
ignore it thing, it's monitored and nudged about periodically. If the
clock was too far off (e.g. several hours, or the wrong date), it'd
refuse to tweak the clock, and you'd have to manually intervene. Of
course, if you weren't watching bootup messages, you wouldn't see any
notification about failure, and you'd only know about it by looking at
your toolbar clock, for instance.
2. At boot time use the NTP daemon to hard force the clock to the same
time as found from a remote server, then have the NTP client keeping it
running correctly on time.
3. Use a NTP utility (e.g. ntpdate) to hard force the clock to the same
time as found from a remote server, then have a NTP daemon keep the
clock running on time.
NTP is one of those odd ones, where the daemon is also a client (it
handles your clock using an external time source) and a server (it can
supply other computers on the network with its time).
[bobg@box10 ~]$ ssh -XC bobg@box48
bobg@box48's password:
Warning: No xauth data; using fake authentication data for X11 forwarding.
Last login: Sun Jan 4 20:36:00 2015 from box10
"fake authentication" What's that about?
No idea, can't say I've seen that before. Sounds like you need to read
up about SSH.
[bobg@box48 ~]$ chronyc sources
210 Number of sources = 1
MS Name/IP address Stratum Poll Reach LastRx Last sample
===============================================================================
^* box10 2 6 377 42 -151us[-1713us]
+/- 321ms
This is the result with only one server, I wanted to see one working
before I confused things with a second unknown. It's interesting to note
that although it is configured to use 192.168.1.10 it correctly
identifies the ntp server as box10 which I assume it gets from /etc/hosts?
That, or DNS, I would say. Like the ntp query program, there's a "-n"
option to stop name lookups. And there is some setting regarding DNS,
mentioned in the documentation.
I found volumes written on ntp and chrony and the remarkable things
it
can do but what I really need is some basic information, primarily how
to cause it to run a new time correction cycle so I can know if a
configuration change has done what I want it to do?
I'd expect looking at a status report, then restarting it might provide
a clue. Such as looking at the time since it last received data, and
how far it is from correct. And, I would have said look at the tail
of /var/log/messages (not sure if that will still work, these days).
Since chronyd is a daemon, I expect it to run in the background and
continually tinker with the clock, like NTP did.
On the other hand, chronyc is a command line program which can be used
to control chrony, and, I think be able to use as a one-shot time
setter.
Try man chrony, run chronyc and type in help, or info chrony. The last
one seems more informative, though hideous to read. There's a
"makestep" chronyc command to force an instant clock set, rather than
gradually slewing the clock into sync.
--
tim@localhost ~]$ uname -rsvp
Linux 3.17.7-200.fc20.i686 #1 SMP Wed Dec 17 04:08:31 UTC 2014 i686
All mail to my mailbox is automatically deleted, there is no point trying
to privately email me, I will only read messages posted to the public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
4:49 a.m.
New subject: chrony lan -_-
On 05.01.2015 11:20, poma wrote:
On 05.01.2015 10:41, Bob Goodwin wrote:
>
> On 01/05/15 01:43, Tim wrote:
>> Allegedly, on or about 04 January 2015, Bob Goodwin sent:
>>>> But I could have two crony/NTP servers no? In addition to box10 I
>>>> could make box7 an ntp server and list both in the clients
>>>> configuration file?
>> Yes, that's actually a good idea. If you already have two PCs
>> consulting the outside world for time, make them servers on your LAN,
>> and let your clients use them. That lets your clients get time from
>> either server, if one of them isn't available for some reason.
>>
>> A NTP client really should consult several NTP servers (some say at
>> least four), and those servers should consult more than one server, too
>> (unless they are*the* master server with an atomic clock, etc). Some
>> math is done to work out the differences/delays between multiple
>> servers, to work out which servers are closer up the chain to a master
>> server, and your own clock's inaccuracies, and your client does its
>> trick as it sees best based on that information.
>
> Crony/ntp appears to be giving the expected results this morning after
> being left to do whatever it does overnight and rebooting the client
> again. So I conclude that although I can do systemctl restart chrony
> that doesn't initiate an update of the system time correction? Not
> knowing how to do that makes it difficult to test ...
>
> [bobg@box10 ~]$ ssh -XC bobg@box48
> bobg@box48's password:
> Warning: No xauth data; using fake authentication data for X11 forwarding.
> Last login: Sun Jan 4 20:36:00 2015 from box10
>
> "fake authentication" What's that about?
>
> [bobg@box48 ~]$ chronyc sources
> 210 Number of sources = 1
> MS Name/IP address Stratum Poll Reach LastRx Last sample
> ===============================================================================
> ^* box10 2 6 377 42 -151us[-1713us]
> +/- 321ms
>
> This is the result with only one server, I wanted to see one working
> before I confused things with a second unknown. It's interesting to note
> that although it is configured to use 192.168.1.10 it correctly
> identifies the ntp server as box10 which I assume it gets from /etc/hosts?
>
> I found volumes written on ntp and chrony and the remarkable things it
> can do but what I really need is some basic information, primarily how
> to cause it to run a new time correction cycle so I can know if a
> configuration change has done what I want it to do?
>
$ system-config-date
The simplest possible docs
$ yelp ghelp:system-config-date
&
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/...
$ man 1 timedatectl
&
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7-...
4:20 a.m.
New subject: chrony lan - OT -_-
On 05.01.2015 10:41, Bob Goodwin wrote:
On 01/05/15 01:43, Tim wrote:
> Allegedly, on or about 04 January 2015, Bob Goodwin sent:
>>> But I could have two crony/NTP servers no? In addition to box10 I
>>> could make box7 an ntp server and list both in the clients
>>> configuration file?
> Yes, that's actually a good idea. If you already have two PCs
> consulting the outside world for time, make them servers on your LAN,
> and let your clients use them. That lets your clients get time from
> either server, if one of them isn't available for some reason.
>
> A NTP client really should consult several NTP servers (some say at
> least four), and those servers should consult more than one server, too
> (unless they are*the* master server with an atomic clock, etc). Some
> math is done to work out the differences/delays between multiple
> servers, to work out which servers are closer up the chain to a master
> server, and your own clock's inaccuracies, and your client does its
> trick as it sees best based on that information.
Crony/ntp appears to be giving the expected results this morning after
being left to do whatever it does overnight and rebooting the client
again. So I conclude that although I can do systemctl restart chrony
that doesn't initiate an update of the system time correction? Not
knowing how to do that makes it difficult to test ...
[bobg@box10 ~]$ ssh -XC bobg@box48
bobg@box48's password:
Warning: No xauth data; using fake authentication data for X11 forwarding.
Last login: Sun Jan 4 20:36:00 2015 from box10
"fake authentication" What's that about?
[bobg@box48 ~]$ chronyc sources
210 Number of sources = 1
MS Name/IP address Stratum Poll Reach LastRx Last sample
===============================================================================
^* box10 2 6 377 42 -151us[-1713us]
+/- 321ms
This is the result with only one server, I wanted to see one working
before I confused things with a second unknown. It's interesting to note
that although it is configured to use 192.168.1.10 it correctly
identifies the ntp server as box10 which I assume it gets from /etc/hosts?
I found volumes written on ntp and chrony and the remarkable things it
can do but what I really need is some basic information, primarily how
to cause it to run a new time correction cycle so I can know if a
configuration change has done what I want it to do?
$ system-config-date
3:41 a.m.
New subject: chrony lan - OT -
On 01/05/15 01:43, Tim wrote:
Allegedly, on or about 04 January 2015, Bob Goodwin sent:
> >But I could have two crony/NTP servers no? In addition to box10 I
> >could make box7 an ntp server and list both in the clients
> >configuration file?
Yes, that's actually a good idea. If you already have two PCs
consulting the outside world for time, make them servers on your LAN,
and let your clients use them. That lets your clients get time from
either server, if one of them isn't available for some reason.
A NTP client really should consult several NTP servers (some say at
least four), and those servers should consult more than one server, too
(unless they are*the* master server with an atomic clock, etc). Some
math is done to work out the differences/delays between multiple
servers, to work out which servers are closer up the chain to a master
server, and your own clock's inaccuracies, and your client does its
trick as it sees best based on that information.
Crony/ntp appears to be giving the expected results this morning after
being left to do whatever it does overnight and rebooting the client
again. So I conclude that although I can do systemctl restart chrony
that doesn't initiate an update of the system time correction? Not
knowing how to do that makes it difficult to test ...
[bobg@box10 ~]$ ssh -XC bobg@box48
bobg@box48's password:
Warning: No xauth data; using fake authentication data for X11 forwarding.
Last login: Sun Jan 4 20:36:00 2015 from box10
"fake authentication" What's that about?
[bobg@box48 ~]$ chronyc sources
210 Number of sources = 1
MS Name/IP address Stratum Poll Reach LastRx Last sample
===============================================================================
^* box10 2 6 377 42 -151us[-1713us]
+/- 321ms
This is the result with only one server, I wanted to see one working
before I confused things with a second unknown. It's interesting to note
that although it is configured to use 192.168.1.10 it correctly
identifies the ntp server as box10 which I assume it gets from /etc/hosts?
I found volumes written on ntp and chrony and the remarkable things it
can do but what I really need is some basic information, primarily how
to cause it to run a new time correction cycle so I can know if a
configuration change has done what I want it to do?
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
12:43 a.m.
New subject: chrony lan - OT -
Allegedly, on or about 04 January 2015, Bob Goodwin sent:
But I could have two crony/NTP servers no? In addition to box10 I
could make box7 an ntp server and list both in the clients
configuration file?
Yes, that's actually a good idea. If you already have two PCs
consulting the outside world for time, make them servers on your LAN,
and let your clients use them. That lets your clients get time from
either server, if one of them isn't available for some reason.
A NTP client really should consult several NTP servers (some say at
least four), and those servers should consult more than one server, too
(unless they are *the* master server with an atomic clock, etc). Some
math is done to work out the differences/delays between multiple
servers, to work out which servers are closer up the chain to a master
server, and your own clock's inaccuracies, and your client does its
trick as it sees best based on that information.
--
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
ZNQR LBH YBBX
8:01 p.m.
New subject: chrony lan - OT -_-
On 04.01.2015 17:29, Bob Goodwin wrote:
On 01/04/15 06:26, poma wrote:
> ACTING AS AN NTP SERVER /usr/share/doc/chrony/chrony.conf.example
>
http://git.tuxfamily.org/chrony/chrony.git/?p=chrony/chrony.git;a=blob;f=...
> Good morning Alfred
On 01/03/15 19:04, Ed Greshko wrote:
> Assuming that one of them is running chronyd you'll need to configure it to Allow
NTP client access from local network using the "allow" directive in the config
file. You'll also need to change the firewall settings to allow incoming ntp requests
as this is normally blocked.
>
> On the SL7 side you'll need to configure it to point to the workstation acting as
the time server instead of servers on the internet.
So it appears that I need to change box10 to make it an ntp server:
# Allow NTP client access from local network.
#allow 192.168/16
allow 192.168.1.0/24
In the firewalld GUI I have checked NTP under SERVICES and made it
PERMANENT. I'm really unsure of myself there!
And then I assume I can add 192.168.1.10 [box10 ntp server] at the top
of the list of the pool of public servers in /etc/chrony.conf in
192.168.1.48 [the samba server to be blocked from the internet]?
Perhaps instead of 192.168.1.10 I could use 192.168.1.0/24?
I would like some reassurance on this ...
Bob
All three combinations should work.
/etc/chrony.conf
...
# Allow NTP client access from local network.
allow 192.168.1
# or
allow 192.168.1/24
# or
allow 192.168.1.2
allow 192.168.1.3
...
$ systemctl restart chronyd.service
Open port 123/udp
/usr/lib/firewalld/services/ntp.xml
$ firewall-cmd --permanent --add-service=ntp
$ firewall-cmd --reload
$ firewall-cmd --query-service=ntp
man 1 firewall-cmd
6:29 p.m.
New subject: chrony lan - OT -
On 01/05/15 07:44, Bob Goodwin wrote:
server 192.168.1.10
Should be...
server 192.168.1.10 iburst
Works here...
[root@f21 etc]# chronyc sources
210 Number of sources = 1
MS Name/IP address Stratum Poll Reach LastRx Last sample
===============================================================================
^* 192.168.1.18 3 6 17 16 -24us[ -85us] +/- 158ms
--
If you can't laugh at yourself, others will gladly oblige.
5:44 p.m.
New subject: chrony lan - OT -
On 01/04/15 18:01, Ed Greshko wrote:
> But I could have two crony/NTP servers no? In addition to box10 I
could make box7 an ntp server and list both in the clients configuration file? That
doesn't seem problematic or am I missing something?
> >
> >
Yes, you can have multiple "server" lines in the config file. And, if you
want, you can use the FQDN and configure your DNS to serve IP addresses in a round-robin
fashion.
With the following config.:
# Use public servers from the pool.ntp.org project.
# Please consider joining the pool (http://www.pool.ntp.org/join.html).
server 192.168.1.10
# server 0.rhel.pool.ntp.org iburst
# server 1.rhel.pool.ntp.org iburst
# server 2.rhel.pool.ntp.org iburst
# server 3.rhel.pool.ntp.org iburst
I see:
[root@box48 bobg]# chronyc sources
210 Number of sources = 1
MS Name/IP address Stratum Poll Reach LastRx Last sample
===============================================================================
^? box10 0 6 0 10y +0ns[ +0ns]
+/- 0ns
Which looks like it's not working? The "?" must mean something?
I've tried disabling firewalld and selinux, neither appears to be
causing a problem.
I've done:
systemctl restart chronyd.service
and then:
[root@box48 bobg]# systemctl status chronyd.service
chronyd.service - NTP client/server
Loaded: loaded (/usr/lib/systemd/system/chronyd.service; enabled)
Active: active (running) since Sun 2015-01-04 18:23:09 EST; 18min ago
Process: 17861 ExecStartPost=/usr/libexec/chrony-helper
add-dhclient-servers (code=exited, status=0/SUCCESS)
Process: 17858 ExecStart=/usr/sbin/chronyd -u chrony $OPTIONS
(code=exited, status=0/SUCCESS)
Main PID: 17860 (chronyd)
CGroup: /system.slice/chronyd.service
└─17860 /usr/sbin/chronyd -u chrony
Jan 04 18:23:09 box48 chronyd[17860]: chronyd version 1.29.1 starting
Jan 04 18:23:09 box48 chronyd[17860]: Linux kernel major=3 minor=10 patch=0
Jan 04 18:23:09 box48 chronyd[17860]: hz=100 shift_hz=7
freq_scale=1.00000000 nominal_tick=10000 slew_delta_tick=833
max_tick_bias=1000 shift_pll=2
Jan 04 18:23:09 box48 chronyd[17860]: Frequency 38.039 +/- 0.463 ppm
read from /var/lib/chrony/drift
Jan 04 18:23:09 box48 systemd[1]: Started NTP client/server.
And I still can't tell from any of this if it's working. I'm in over my
head here ...
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
5:01 p.m.
New subject: chrony lan - OT -
On 01/05/15 03:35, Bob Goodwin wrote:
On 01/04/15 14:24, Ed Greshko wrote:
>> >So it appears that I need to change box10 to make it an ntp server:
>> >
>> ># Allow NTP client access from local network.
>> >#allow 192.168/16
>> >allow 192.168.1.0/24
>> >
>> >In the firewalld GUI I have checked NTP under SERVICES and made it PERMANENT.
I'm really unsure of myself there!
>> >
>> >And then I assume I can add 192.168.1.10 [box10 ntp server] at the top of the
list of the pool of public servers in /etc/chrony.conf in 192.168.1.48 [the samba server
to be blocked from the internet]?
>> >
>> >Perhaps instead of 192.168.1.10 I could use 192.168.1.0/24?
>> >
>> >I would like some reassurance on this ...
>> >
> You want the actual IP address of the ntp server. It isn't a case of searching.
It is a case of knowing.
>
But I could have two crony/NTP servers no? In addition to box10 I could make box7 an ntp
server and list both in the clients configuration file? That doesn't seem problematic
or am I missing something?
Yes, you can have multiple "server" lines in the config file. And, if you want,
you can use the FQDN and configure your DNS to serve IP addresses in a round-robin
fashion.
--
If you can't laugh at yourself, others will gladly oblige.
5:26 a.m.
New subject: chrony lan - OT -
On 04.01.2015 10:17, Bob Goodwin wrote:
On 01/04/15 04:14, Bob Goodwin wrote:
>
> On 01/03/15 20:55, g wrote:
>> laughing at yourself or using syncing servers to a workstation.;-)
Sorry, I somehow got that message addressed wrong!
Bob
ACTING AS AN NTP SERVER
/usr/share/doc/chrony/chrony.conf.example
http://git.tuxfamily.org/chrony/chrony.git/?p=chrony/chrony.git;a=blob;f=...
Good morning Alfred
1:35 p.m.
New subject: chrony lan - OT -
On 01/04/15 14:24, Ed Greshko wrote:
> >So it appears that I need to change box10 to make it an ntp
server:
> >
> ># Allow NTP client access from local network.
> >#allow 192.168/16
> >allow 192.168.1.0/24
> >
> >In the firewalld GUI I have checked NTP under SERVICES and made it PERMANENT.
I'm really unsure of myself there!
> >
> >And then I assume I can add 192.168.1.10 [box10 ntp server] at the top of the
list of the pool of public servers in /etc/chrony.conf in 192.168.1.48 [the samba server
to be blocked from the internet]?
> >
> >Perhaps instead of 192.168.1.10 I could use 192.168.1.0/24?
> >
> >I would like some reassurance on this ...
> >
You want the actual IP address of the ntp server. It isn't a case of searching. It
is a case of knowing.
But I could have two crony/NTP servers no? In addition to box10 I could
make box7 an ntp server and list both in the clients configuration file?
That doesn't seem problematic or am I missing something?
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
1:24 p.m.
New subject: chrony lan - OT -
On 01/05/15 00:29, Bob Goodwin wrote:
On 01/04/15 06:26, poma wrote:
> ACTING AS AN NTP SERVER /usr/share/doc/chrony/chrony.conf.example
http://git.tuxfamily.org/chrony/chrony.git/?p=chrony/chrony.git;a=blob;f=...
Good morning Alfred
On 01/03/15 19:04, Ed Greshko wrote:
> Assuming that one of them is running chronyd you'll need to configure it to Allow
NTP client access from local network using the "allow" directive in the config
file. You'll also need to change the firewall settings to allow incoming ntp requests
as this is normally blocked.
>
> On the SL7 side you'll need to configure it to point to the workstation acting as
the time server instead of servers on the internet.
So it appears that I need to change box10 to make it an ntp server:
# Allow NTP client access from local network.
#allow 192.168/16
allow 192.168.1.0/24
In the firewalld GUI I have checked NTP under SERVICES and made it PERMANENT. I'm
really unsure of myself there!
And then I assume I can add 192.168.1.10 [box10 ntp server] at the top of the list of the
pool of public servers in /etc/chrony.conf in 192.168.1.48 [the samba server to be blocked
from the internet]?
Perhaps instead of 192.168.1.10 I could use 192.168.1.0/24?
I would like some reassurance on this ...
You want the actual IP address of the ntp server. It isn't a case of searching. It
is a case of knowing.
--
If you can't laugh at yourself, others will gladly oblige.
10:29 a.m.
New subject: chrony lan - OT -
On 01/04/15 06:26, poma wrote:
ACTING AS AN NTP SERVER /usr/share/doc/chrony/chrony.conf.example
http://git.tuxfamily.org/chrony/chrony.git/?p=chrony/chrony.git;a=blob;f=...
Good morning Alfred
On 01/03/15 19:04, Ed Greshko wrote:
Assuming that one of them is running chronyd you'll need to
configure it to Allow NTP client access from local network using the "allow"
directive in the config file. You'll also need to change the firewall settings to
allow incoming ntp requests as this is normally blocked.
On the SL7 side you'll need to configure it to point to the workstation acting as the
time server instead of servers on the internet.
So it appears that I need to change box10 to make it an ntp server:
# Allow NTP client access from local network.
#allow 192.168/16
allow 192.168.1.0/24
In the firewalld GUI I have checked NTP under SERVICES and made it
PERMANENT. I'm really unsure of myself there!
And then I assume I can add 192.168.1.10 [box10 ntp server] at the top
of the list of the pool of public servers in /etc/chrony.conf in
192.168.1.48 [the samba server to be blocked from the internet]?
Perhaps instead of 192.168.1.10 I could use 192.168.1.0/24?
I would like some reassurance on this ...
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
3:17 a.m.
New subject: chrony lan - OT -
On 01/04/15 04:14, Bob Goodwin wrote:
On 01/03/15 20:55, g wrote:
> laughing at yourself or using syncing servers to a workstation.;-)
Sorry, I
somehow got that message addressed wrong!
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
7:55 p.m.
On 01/03/2015 06:42 PM, Bob Goodwin wrote:
On 01/03/15 19:04, Ed Greshko wrote:
<<>>
> So, you need a local ntp server.....
>
> You have 2 workstations getting their time sync from the internet
> as their access is enabled. One of those could act as a time
> server.
<<>>
> -- If you can't laugh at yourself, others will gladly
oblige.
Good, I will give this a try tomorrow
laughing at yourself or using syncing servers to a workstation. ;-)
hi Bob,
along with what Ed Greshko suggested,,,
being that you have mentioned the internet time cost with being on
satellite, you prefer to limit running ntp for clock syncing.
so what is wrong with WWV? surely you can receive WWV on 10 meters from
where you are.
and if so, you can tweak 'drift rate' /etc/adjtime. just check WWV time
every 24 hrs at same time for a week and tweak 'drift rate' to compensate.
after 7 days, check every other day until you get drift to minimal.
lines in /etc/adjtime are;
line 1 numbers:
1st= systematic drift rate in seconds per day,
floating point decimal.
2nd= resulting number of seconds since 1969 UTC of most recent
adjustment or calibration, decimal integer.
3rd= zero, for compatibility with clock(8), decimal integer.
line 2 number:
resulting number of seconds since 1969 UTC of most recent
calibration. zero if there has been no calibration yet
or it is known that any previous calibration is moot.
ie, because the hardware clock has been found, since that
calibration, not to contain a valid time, decimal integer.
line 3:
"UTC" or "LOCAL". setting of the hardware.
you could also use 'hwclock' to adjust /etc/adjtime. see "man hwclock"
for how to use. also, a good article;
http://www.sanfoundry.com/5-hwclock-command-usage-examples-linux/
--
Ευτυχισμένα γενέθλια Χριστού.
happy birthday Jesus.
καλά χριστούγεννα. ευτυχές το νέο έτος.
Merry Christmas. Happy New Year.
peace out.
in a world with out fences, who needs gates.
CentOS GNU/Linux 6.6
tc,hago.
g
.
4:21 p.m.
On 01/03/15 16:39, Ed Greshko wrote:
On 01/04/15 01:27, Bob Goodwin wrote:
> If I block WAN connection to a server will it find a time reference on the LAN?
There's usually a workstation connected and getting it's clock updated but it
looks like the server is not? I haven't found the answer in the man pages ... I was
looking for a configuration setting perhaps.
The configuration file for chronyd can be found by consulting the man page for chronyd.
If no configuration commands are specified on the command line, chronyd
will read the commands from the configuration file (default
/etc/chrony.conf).
Yes, I've looked at that but it's not clear to me.
# Allow NTP client access from local network.
#allow 192.168/16
Will uncommenting this allow the server to find a time sync signal on
the LAN? Even the router keeps time so it seems there should be a way
other than keeping the server connected to the internet for time alone?
It seems I should be able to tell it to get it's time from 192.168.1.1
[the router].
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
6:42 p.m.
On 01/03/15 19:04, Ed Greshko wrote:
On 01/04/15 07:30, Bob Goodwin wrote:
> >
> >On 01/03/15 17:53, Ed Greshko wrote:
>> >>My read of your request is this....
>> >>
>> >>You have a system for which WAN access is blocked. So it can't
contact an ntp server on outside of your LAN for time synchronization.
>> >>
>> >>In that case,*if* you have a local system server as a time source you
need to set the "server" parameter in the config file to tell chronyd what time
sources to contact. If your router can also act as an ntp server you can point to it.
> >I have a LAN with a router and two workstations that get their time from their
internet connection. I have two servers that I prefer not to connect to the internet since
it serves no purpose other than keep the clocks synced. It's a simple matter to block
the server connection in the router.
> >
> >One of the servers is using Scientific Linux 7. It appeared the clock was not
being synced until I opened it's internet access. I would like to fix that ...
> >
So, you need a local ntp server.....
You have 2 workstations getting their time sync from the internet as their access is
enabled. One of those could act as a time server.
Assuming that one of them is running chronyd you'll need to configure it to Allow NTP
client access from local network using the "allow" directive in the config file.
You'll also need to change the firewall settings to allow incoming ntp requests as
this is normally blocked.
On the SL7 side you'll need to configure it to point to the workstation acting as the
time server instead of servers on the internet.
-- If you can't laugh at yourself, others will gladly oblige.
Good, I will give
this a try tomorrow
Thank you Ed.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
6:04 p.m.
On 01/04/15 07:30, Bob Goodwin wrote:
On 01/03/15 17:53, Ed Greshko wrote:
> My read of your request is this....
>
> You have a system for which WAN access is blocked. So it can't contact an ntp
server on outside of your LAN for time synchronization.
>
> In that case,*if* you have a local system server as a time source you need to set
the "server" parameter in the config file to tell chronyd what time sources to
contact. If your router can also act as an ntp server you can point to it.
I have a LAN with a router and two workstations that get their time from their internet
connection. I have two servers that I prefer not to connect to the internet since it
serves no purpose other than keep the clocks synced. It's a simple matter to block the
server connection in the router.
One of the servers is using Scientific Linux 7. It appeared the clock was not being
synced until I opened it's internet access. I would like to fix that ...
So, you need a local ntp server.....
You have 2 workstations getting their time sync from the internet as their access is
enabled. One of those could act as a time server.
Assuming that one of them is running chronyd you'll need to configure it to Allow NTP
client access from local network using the "allow" directive in the config file.
You'll also need to change the firewall settings to allow incoming ntp requests as
this is normally blocked.
On the SL7 side you'll need to configure it to point to the workstation acting as the
time server instead of servers on the internet.
--
If you can't laugh at yourself, others will gladly oblige.
5:30 p.m.
On 01/03/15 17:53, Ed Greshko wrote:
My read of your request is this....
You have a system for which WAN access is blocked. So it can't contact an ntp server
on outside of your LAN for time synchronization.
In that case,*if* you have a local system server as a time source you need to set the
"server" parameter in the config file to tell chronyd what time sources to
contact. If your router can also act as an ntp server you can point to it.
I have a
LAN with a router and two workstations that get their time from
their internet connection. I have two servers that I prefer not to
connect to the internet since it serves no purpose other than keep the
clocks synced. It's a simple matter to block the server connection in
the router.
One of the servers is using Scientific Linux 7. It appeared the clock
was not being synced until I opened it's internet access. I would like
to fix that ...
--
http://www.qrz.com/db/W2BOD
box10 Fedora-21/64bit Linux/XFCE
4:53 p.m.
On 01/04/15 06:21, Bob Goodwin wrote:
On 01/03/15 16:39, Ed Greshko wrote:
> On 01/04/15 01:27, Bob Goodwin wrote:
>> If I block WAN connection to a server will it find a time reference on the LAN?
There's usually a workstation connected and getting it's clock updated but it
looks like the server is not? I haven't found the answer in the man pages ... I was
looking for a configuration setting perhaps.
> The configuration file for chronyd can be found by consulting the man page for
chronyd.
>
> If no configuration commands are specified on the command line, chronyd
> will read the commands from the configuration file (default
> /etc/chrony.conf).
>
>
>
Yes, I've looked at that but it's not clear to me.
# Allow NTP client access from local network.
#allow 192.168/16
Will uncommenting this allow the server to find a time sync signal on the LAN? Even the
router keeps time so it seems there should be a way other than keeping the server
connected to the internet for time alone? It seems I should be able to tell it to get
it's time from 192.168.1.1 [the router].
My read of your request is this....
You have a system for which WAN access is blocked. So it can't contact an ntp server
on outside of your LAN for time synchronization.
In that case, *if* you have a local system server as a time source you need to set the
"server" parameter in the config file to tell chronyd what time sources to
contact. If your router can also act as an ntp server you can point to it.
--
If you can't laugh at yourself, others will gladly oblige.
3:39 p.m.
On 01/04/15 01:27, Bob Goodwin wrote:
If I block WAN connection to a server will it find a time reference
on the LAN? There's usually a workstation connected and getting it's clock updated
but it looks like the server is not? I haven't found the answer in the man pages ... I
was looking for a configuration setting perhaps.
The configuration file for chronyd can be found by consulting the man page for chronyd.
If no configuration commands are specified on the command line, chronyd
will read the commands from the configuration file (default
/etc/chrony.conf).
--
If you can't laugh at yourself, others will gladly oblige.
3406
days inactive
3408
days old
22 comments
5 participants
participants (5)
-
Bob Goodwin -
Ed Greshko -
g -
poma -
Tim