[gentoo-user] ntp won't synchronize

[gentoo-user] ntp won't synchronize

[matthew.garman]

I've been struggling with ntp for some time now.  I've followed the
gentoo wiki HOWTO for ntp:

    http://gentoo-wiki.com/HOWTO_Setup_MythTV#NTP

As well as many other sources over the months.  Basically, ntpq
shows that I am not synchronized to any peers:

# ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 Time4.Stupi.SE  .PPS.            1 u   37   64  377  124.516  -11202. 1225.91
 thesimonet.org  .TRUE.           1 u   38   64  377  109.329  -11213. 1232.51
 fin.rshell.net  192.114.62.249   3 u   32   64  377  205.910  -9569.1 1521.31
 titan.cais.rnp. 32.233.177.224   2 u   44   64  377  165.240  -11991. 1915.26

That first space in the peers list is a tally mark that, according to
the ntpq documentation at:

    http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html

Means that those peers are all rejected: "The peer is discarded as
unreachable, synchronized to this server (synch loop) or outrageous
synchronization distance."

Likewise:
# ntpq -c rv | grep stratum
processor="i686", system="Linux/2.6.14-gentoo-r5", leap=11, stratum=16,

I should be at stratum 3, not 16 (which means I'm not synchronized
to anything).

My /etc/ntp.conf:

tinker panic 0
minpoll 4
maxpoll 10
server pool.ntp.org
server 0.pool.ntp.org
server 1.pool.ntp.org
server 2.pool.ntp.org
driftfile	/var/lib/ntp/ntp.drift
logfile	/var/log/ntp.log
restrict default nomodify nopeer
restrict 127.0.0.1


Nothing special there!  To further clarify, I actually have this
problem on *two* machines.  FWIW, I've also tried the OpenBSD
OpenNTP package, but haven't had any luck with that either (except
on my OpenBSD machine!).  Using OpenNTP on these two machines
*seemed* to work fine, but the computers still gained time too
quickly.

I've been fighting this for what seems like forever.  If anyone has
any insight or thoughts, I'm happy to hear it!

Thanks,
Matt

-- 
Matt Garman
email at: http://raw-sewage.net/index.php?file=email
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] ntp won't synchronize

[Francesco Riosa]

matthew.garman@gmail.com wrote:
> I've been struggling with ntp for some time now.  I've followed the
> gentoo wiki HOWTO for ntp:
>
>     http://gentoo-wiki.com/HOWTO_Setup_MythTV#NTP
>
> As well as many other sources over the months.  Basically, ntpq
> shows that I am not synchronized to any peers:
>
> # ntpq -p
>      remote           refid      st t when poll reach   delay   offset  jitter
> ==============================================================================
>  Time4.Stupi.SE  .PPS.            1 u   37   64  377  124.516  -11202. 1225.91
>  thesimonet.org  .TRUE.           1 u   38   64  377  109.329  -11213. 1232.51
>  fin.rshell.net  192.114.62.249   3 u   32   64  377  205.910  -9569.1 1521.31
>  titan.cais.rnp. 32.233.177.224   2 u   44   64  377  165.240  -11991. 1915.26
>
> That first space in the peers list is a tally mark that, according to
> the ntpq documentation at:
>
>     http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html
>
> Means that those peers are all rejected: "The peer is discarded as
> unreachable, synchronized to this server (synch loop) or outrageous
> synchronization distance."
>
> Likewise:
> # ntpq -c rv | grep stratum
> processor="i686", system="Linux/2.6.14-gentoo-r5", leap=11, stratum=16,
>
> I should be at stratum 3, not 16 (which means I'm not synchronized
> to anything).
>
> My /etc/ntp.conf:
>
> tinker panic 0
> minpoll 4
> maxpoll 10
> server pool.ntp.org
> server 0.pool.ntp.org
> server 1.pool.ntp.org
> server 2.pool.ntp.org
> driftfile	/var/lib/ntp/ntp.drift
> logfile	/var/log/ntp.log
> restrict default nomodify nopeer
> restrict 127.0.0.1
>
>
> Nothing special there!  To further clarify, I actually have this
> problem on *two* machines.  FWIW, I've also tried the OpenBSD
> OpenNTP package, but haven't had any luck with that either (except
> on my OpenBSD machine!).  Using OpenNTP on these two machines
> *seemed* to work fine, but the computers still gained time too
> quickly.
>
> I've been fighting this for what seems like forever.  If anyone has
> any insight or thoughts, I'm happy to hear it!
>
> Thanks,
> Matt
>
>   

mine:

$ ntpq -p
     remote           refid      st t when poll reach   delay   offset 
jitter
==============================================================================
+ns.pnpitalia.it 62.101.81.203    3 u  764 1024  377    0.242    7.403  
4.093
*192.168.76.2     193.204.114.105  2 u  215 1024  377    1.170   11.839 
39.537

$ cat /etc/ntp.conf
# Generated by dhcpcd for interface eth1
restrict default noquery notrust nomodify
restrict 127.0.0.1
restrict 192.168.76.1 nomodify notrap noquery
server 192.168.76.1
restrict 192.168.76.2 nomodify notrap noquery
server 192.168.76.2
driftfile /var/lib/ntp/ntp.drift
logfile /var/log/ntp.log

^^^ this is to join a _local_ server

I'm a bit rust on this but should the "restrict" parameters allow
servers from pool.ntp.org ?
also if there is delta time too big it wont syncronize , before to start
the ntp server try a

ntpdate -4b 0.pool.ntp.org
ntpdate -4b 1.pool.ntp.org

then start ntp, leave pass few hours and then check again the result of
ntpq -p

as said I'm a bit rush but small variation of what described here should
work.


-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] ntp won't synchronize

[Devon Miller]

[-- Attachment #1: Type: text/plain, Size: 2483 bytes --]

Make sure you have told you firewall to allow port 123 for both TCP & UDP.
I had the same behavior until I did that.

dcm

On 1/9/06, matthew.garman@gmail.com <matthew.garman@gmail.com> wrote:
>
>
> I've been struggling with ntp for some time now.  I've followed the
> gentoo wiki HOWTO for ntp:
>
>     http://gentoo-wiki.com/HOWTO_Setup_MythTV#NTP
>
> As well as many other sources over the months.  Basically, ntpq
> shows that I am not synchronized to any peers:
>
> # ntpq -p
>      remote           refid      st t when poll reach   delay
> offset  jitter
>
> ==============================================================================
> Time4.Stupi.SE  .PPS.            1 u   37   64  377  124.516  -11202.
> 1225.91
> thesimonet.org  .TRUE.           1 u   38   64  377  109.329  -11213.
> 1232.51
> fin.rshell.net  192.114.62.249   3 u   32   64  377  205.910  -9569.1
> 1521.31
> titan.cais.rnp. 32.233.177.224   2 u   44   64  377  165.240  -11991.
> 1915.26
>
> That first space in the peers list is a tally mark that, according to
> the ntpq documentation at:
>
>     http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html
>
> Means that those peers are all rejected: "The peer is discarded as
> unreachable, synchronized to this server (synch loop) or outrageous
> synchronization distance."
>
> Likewise:
> # ntpq -c rv | grep stratum
> processor="i686", system="Linux/2.6.14-gentoo-r5", leap=11, stratum=16,
>
> I should be at stratum 3, not 16 (which means I'm not synchronized
> to anything).
>
> My /etc/ntp.conf:
>
> tinker panic 0
> minpoll 4
> maxpoll 10
> server pool.ntp.org
> server 0.pool.ntp.org
> server 1.pool.ntp.org
> server 2.pool.ntp.org
> driftfile       /var/lib/ntp/ntp.drift
> logfile /var/log/ntp.log
> restrict default nomodify nopeer
> restrict 127.0.0.1
>
>
> Nothing special there!  To further clarify, I actually have this
> problem on *two* machines.  FWIW, I've also tried the OpenBSD
> OpenNTP package, but haven't had any luck with that either (except
> on my OpenBSD machine!).  Using OpenNTP on these two machines
> *seemed* to work fine, but the computers still gained time too
> quickly.
>
> I've been fighting this for what seems like forever.  If anyone has
> any insight or thoughts, I'm happy to hear it!
>
> Thanks,
> Matt
>
> --
> Matt Garman
> email at: http://raw-sewage.net/index.php?file=email
> --
> gentoo-user@gentoo.org mailing list
>
>

[-- Attachment #2: Type: text/html, Size: 4074 bytes --]

Re: [gentoo-user] ntp won't synchronize

[Devon Miller]

[-- Attachment #1: Type: text/plain, Size: 2483 bytes --]

Make sure you have told you firewall to allow port 123 for both TCP & UDP.
I had the same behavior until I did that.

dcm

On 1/9/06, matthew.garman@gmail.com <matthew.garman@gmail.com> wrote:
>
>
> I've been struggling with ntp for some time now.  I've followed the
> gentoo wiki HOWTO for ntp:
>
>     http://gentoo-wiki.com/HOWTO_Setup_MythTV#NTP
>
> As well as many other sources over the months.  Basically, ntpq
> shows that I am not synchronized to any peers:
>
> # ntpq -p
>      remote           refid      st t when poll reach   delay
> offset  jitter
>
> ==============================================================================
> Time4.Stupi.SE  .PPS.            1 u   37   64  377  124.516  -11202.
> 1225.91
> thesimonet.org  .TRUE.           1 u   38   64  377  109.329  -11213.
> 1232.51
> fin.rshell.net  192.114.62.249   3 u   32   64  377  205.910  -9569.1
> 1521.31
> titan.cais.rnp. 32.233.177.224   2 u   44   64  377  165.240  -11991.
> 1915.26
>
> That first space in the peers list is a tally mark that, according to
> the ntpq documentation at:
>
>     http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html
>
> Means that those peers are all rejected: "The peer is discarded as
> unreachable, synchronized to this server (synch loop) or outrageous
> synchronization distance."
>
> Likewise:
> # ntpq -c rv | grep stratum
> processor="i686", system="Linux/2.6.14-gentoo-r5", leap=11, stratum=16,
>
> I should be at stratum 3, not 16 (which means I'm not synchronized
> to anything).
>
> My /etc/ntp.conf:
>
> tinker panic 0
> minpoll 4
> maxpoll 10
> server pool.ntp.org
> server 0.pool.ntp.org
> server 1.pool.ntp.org
> server 2.pool.ntp.org
> driftfile       /var/lib/ntp/ntp.drift
> logfile /var/log/ntp.log
> restrict default nomodify nopeer
> restrict 127.0.0.1
>
>
> Nothing special there!  To further clarify, I actually have this
> problem on *two* machines.  FWIW, I've also tried the OpenBSD
> OpenNTP package, but haven't had any luck with that either (except
> on my OpenBSD machine!).  Using OpenNTP on these two machines
> *seemed* to work fine, but the computers still gained time too
> quickly.
>
> I've been fighting this for what seems like forever.  If anyone has
> any insight or thoughts, I'm happy to hear it!
>
> Thanks,
> Matt
>
> --
> Matt Garman
> email at: http://raw-sewage.net/index.php?file=email
> --
> gentoo-user@gentoo.org mailing list
>
>

[-- Attachment #2: Type: text/html, Size: 4074 bytes --]

[gentoo-user] Re: ntp won't synchronize

[James]

Devon Miller <devon.c.miller <at> gmail.com> writes:

> Make sure you have told you firewall to allow port 123 for both TCP &
> UDP.I had the same behavior until I did that.dcm

Well my firewall should allow outgoing initiated sessions from the
ntpd (internal) server. From what I read, the remote ntpd server 
does not initiate communications, it my server that initates the
communications?

That assumed, my firewall rules (which are undergoing revision) are ok?

OK, I'm new to ntpd, but it looks straightforward.

after emerging, I started up ntpd and added it to the default run level.
Then I checked and got:

ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 ecoca.eed.usv.r 80.96.120.253    2 u    8   64    1  202.639  1436722   0.001

then:
# date
Wed Feb  1 14:58:04 Local time zone must be set--see zic manual page 2006


/etc/localtime is a lock file so where do I correct/set the time zone to
EST (eastern standard time) so ntpd will see it can correct my actual time
(it's off by a little more than (1) hour.

But now when I run 'ntpq -p' I get:
ntpq: read: Connection refused

What did I mess up or miss?


James







man zic and man ntpd did not help, or I missed the file to edit or the 
correct command syntax....

James



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[John Jolet]

On Feb 1, 2006, at 1:25 PM, James wrote:

> Devon Miller <devon.c.miller <at> gmail.com> writes:
>
>> Make sure you have told you firewall to allow port 123 for both TCP &
>> UDP.I had the same behavior until I did that.dcm
>
> Well my firewall should allow outgoing initiated sessions from the
> ntpd (internal) server. From what I read, the remote ntpd server
> does not initiate communications, it my server that initates the
> communications?
>
> That assumed, my firewall rules (which are undergoing revision) are  
> ok?
>
> OK, I'm new to ntpd, but it looks straightforward.
>
> after emerging, I started up ntpd and added it to the default run  
> level.
> Then I checked and got:
>
> ntpq -p
>      remote           refid      st t when poll reach   delay    
> offset  jitter
> ====================================================================== 
> ========
>  ecoca.eed.usv.r 80.96.120.253    2 u    8   64    1  202.639   
> 1436722   0.001
>
> then:
> # date
> Wed Feb  1 14:58:04 Local time zone must be set--see zic manual  
> page 2006
>
>
> /etc/localtime is a lock file so where do I correct/set the time  
> zone to
> EST (eastern standard time) so ntpd will see it can correct my  
> actual time
> (it's off by a little more than (1) hour.
>
> But now when I run 'ntpq -p' I get:
> ntpq: read: Connection refused
>
> What did I mess up or miss?
>
is ntpd dying?  ps -elf|grep ntp should show you something besides  
the grep.
>
> James
>
>
>
>
>
>
>
> man zic and man ntpd did not help, or I missed the file to edit or the
> correct command syntax....
>
> James
>
>
>
> -- 
> gentoo-user@gentoo.org mailing list
>

-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: ntp won't synchronize

[James]

John Jolet <john <at> jolet.net> writes:



> > But now when I run 'ntpq -p' I get:
> > ntpq: read: Connection refused

> is ntpd dying?  ps -elf|grep ntp should show you something besides  
> the grep.

Yep. Attempt stop it and start it again: /etc/init.d/ntpd start 
fails.

Ideas?

James



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[Dave Nebinger]

James wrote:
>> is ntpd dying?  ps -elf|grep ntp should show you something besides  
>> the grep.
>>     
>
> Yep. Attempt stop it and start it again: /etc/init.d/ntpd start 
> fails.
>   

"/etc/init.d/ntpd zap" to clear out the invalid status, then do the 
'start' again.


-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: ntp won't synchronize

[James]

James <wireless <at> tampabay.rr.com> writes:

> 
> John Jolet <john <at> jolet.net> writes:
> 
> > > But now when I run 'ntpq -p' I get:
> > > ntpq: read: Connection refused
> 
> > is ntpd dying?  ps -elf|grep ntp should show you something besides  
> > the grep.
> 
> Yep. Attempt stop it and start it again: /etc/init.d/ntpd start 
> fails.

Well I rebooted and it is running:
9480 ?        SLs    0:00 /usr/sbin/ntpd -p /var/run/ntpd.pid -u ntp:ntp

It corrected the minutes but it's off by exactly one hour. It needs to be
EST (eastern standard time) NY  (Tampa Florida).

so what file do I edit to correct utc to est ?

#date
Wed Feb  1 16:16:51 UTC 2006



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[Richard Fish]

On 2/1/06, James <wireless@tampabay.rr.com> wrote:
> It corrected the minutes but it's off by exactly one hour. It needs to be
> EST (eastern standard time) NY  (Tampa Florida).
>
> so what file do I edit to correct utc to est ?

Try timezone US/Eastern instead of EST.

ln -sf /usr/share/zoneinfo/US/Eastern /etc/localtime

-Richard

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[John Jolet]

On Feb 1, 2006, at 2:18 PM, James wrote:

> James <wireless <at> tampabay.rr.com> writes:
>
>>
>> John Jolet <john <at> jolet.net> writes:
>>
>>>> But now when I run 'ntpq -p' I get:
>>>> ntpq: read: Connection refused
>>
>>> is ntpd dying?  ps -elf|grep ntp should show you something besides
>>> the grep.
>>
>> Yep. Attempt stop it and start it again: /etc/init.d/ntpd start
>> fails.
>
> Well I rebooted and it is running:
> 9480 ?        SLs    0:00 /usr/sbin/ntpd -p /var/run/ntpd.pid -u  
> ntp:ntp
>
> It corrected the minutes but it's off by exactly one hour. It needs  
> to be
> EST (eastern standard time) NY  (Tampa Florida).
>
> so what file do I edit to correct utc to est ?
>
> #date
> Wed Feb  1 16:16:51 UTC 2006
>
in /usr/share/zoneinfo are your timezone files.  copy or symlink the  
correct one to /etc/localtime.
>
>
> -- 
> gentoo-user@gentoo.org mailing list
>

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[W.Kenworthy]

Default behaviour is that if the time zone is off by more than a preset
amount (I think 128ms) it will refuse to sync, and silently fails.

Read the docs for the config file command "tinker panic 0" (and its
implications) which will remove the limitation and allow stepping to the
new time.

Also ntpd can be started with the -g arg to get the same effect.


BillK


On Wed, 2006-02-01 at 19:25 +0000, James wrote:
> Devon Miller <devon.c.miller <at> gmail.com> writes:
> 
> > Make sure you have told you firewall to allow port 123 for both TCP &
> > UDP.I had the same behavior until I did that.dcm
> 
> Well my firewall should allow outgoing initiated sessions from the
> ntpd (internal) server. From what I read, the remote ntpd server 
> does not initiate communications, it my server that initates the
> communications?
> 
> That assumed, my firewall rules (which are undergoing revision) are ok?
> 
> OK, I'm new to ntpd, but it looks straightforward.
> 
> after emerging, I started up ntpd and added it to the default run level.
> Then I checked and got:
> 
> ntpq -p
>      remote           refid      st t when poll reach   delay   offset  jitter
> ==============================================================================
>  ecoca.eed.usv.r 80.96.120.253    2 u    8   64    1  202.639  1436722   0.001
> 
> then:
> # date
> Wed Feb  1 14:58:04 Local time zone must be set--see zic manual page 2006
> 
> 
> /etc/localtime is a lock file so where do I correct/set the time zone to
> EST (eastern standard time) so ntpd will see it can correct my actual time
> (it's off by a little more than (1) hour.
> 
> But now when I run 'ntpq -p' I get:
> ntpq: read: Connection refused
> 
> What did I mess up or miss?
> 
> 
> James
> 
> 
> 
> 
> 
> 
> 
> man zic and man ntpd did not help, or I missed the file to edit or the 
> correct command syntax....
> 
> James
> 
> 
> 
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: ntp won't synchronize

[matthew.garman]

On Wed, Feb 01, 2006 at 07:25:35PM +0000, James wrote:
> ntpq -p
>      remote           refid      st t when poll reach   delay   offset  jitter
> ==============================================================================
>  ecoca.eed.usv.r 80.96.120.253    2 u    8   64    1  202.639  1436722   0.001

You may already know this, but according to my interpretation of the
NTP documentation, the above line says that you are actually NOT
sync'ed to any NTP server.

That first space in the peers list is a tally mark that, according
to the ntpq documentation at:

    http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html

Means that those peers are all rejected: "The peer is discarded as
unreachable, synchronized to this server (synch loop) or outrageous
synchronization distance."

I pointed this out to a friend, who didn't even realize that he
wasn't sync'ing.  I'd bet some real money that many folks are
running ntp, but oblivious to the fact that they actually arent
synchronizing to anyone else.

I don't think using a space as a tic mark to indicate an error (or
at least undesireable) condition is particularly
helpful---especially with no heading!

Matt

-- 
Matt Garman
email at: http://raw-sewage.net/index.php?file=email
-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: ntp won't synchronize

[Harry Putnam]

matthew.garman@gmail.com writes:

> On Wed, Feb 01, 2006 at 07:25:35PM +0000, James wrote:
>> ntpq -p
>>      remote           refid      st t when poll reach   delay   offset  jitter
>> ==============================================================================
>>  ecoca.eed.usv.r 80.96.120.253    2 u    8   64    1  202.639  1436722   0.001

This won't help the above problem but it might help the overall situation.
As I recall ntpd will not adjust the time above a certain threshold.
I don't recall the threshold but its pretty low.

I've found if I make ntp-client to run at boot time it sets the clock
before ntpd gets to it and makes sure its below that threshold.  Once
the clock is below that threshold ntpd will keep it right (assuming
its connecting with time servers)

Throwing in my own question here.  I'd never run ntpq -p before but
now I did and now wondering what a dash before severname means.

     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
[...]
-atlas.jtan.com  128.4.40.12      3 u  488 1024  377   68.824  -11.85987.611
[...]


I couldn't find any reference to it in man ntpq which is loaded with
info.

It refers the reader like this:
 For examples and usage, see the NTP Debugging Techniques page.

But doesn't bother saying what that is supposed to be a page of.

Also found this in there but its so cryptic I couldn't tell if this is
what I was looking for or what it shows a dash being used at least.

 -  outlyer
              The peer is discarded by the clustering algorithm as an outlyer.


-- 
gentoo-user@gentoo.org mailing list