Time error 0x41

On Debian 10, ntpd 4.2.8p12@1.3728-o fails to sync with following error:

kernel reports TIME_ERROR: 0x41: Clock Unsynchronize

here’s ntp.conf:

disable monitor

statsdir /var/log/ntpstats

restrict -4 default kod nomodify notrap nopeer noquery
restrict -6 default kod nomodify notrap nopeer noquery
restrict 127.0.0.1
restrict ::1

server 0.us.pool.ntp.org iburst
server 1.us.pool.ntp.org iburst
server 2.us.pool.ntp.org iburst
server 3.us.pool.ntp.org iburst

server   127.127.1.0
fudge    127.127.1.0 stratum 10
restrict 127.127.1.0

driftfile /var/lib/ntp/drift

ntpq -c sysinfo:

associd=0 status=0614 leap_none, sync_ntp, 1 event, freq_mode,
system peer:        50-205-57-38-static.hfc.comcastbusiness.net:123
system peer mode:   client
leap indicator:     00
stratum:            2
log2 precision:     -23
root delay:         70.634
root dispersion:    3.569
reference ID:       50.205.57.38
reference time:     e3a0c049.c39d770a  Wed, Jan  6 2021 23:03:37.764
system jitter:      0.723169
clock jitter:       1.177
clock wander:       0.000
broadcast delay:    -50.000
symm. auth. delay:  0.000

ntpq -c lpeers:

     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 LOCAL(0)        .LOCL.          10 l  286   64   20    0.000    0.000   0.000
*50-205-57-38-st .GPS.            1 u   19   64   37   70.631    1.618   1.843
-ns1.backplanedn 173.162.192.156  2 u   14   64   37   84.235   -1.575   2.852
+c-73-239-136-18 74.6.168.73      3 u   11   64   37   48.606    1.598   2.522
+time-d.bbnx.net 252.74.143.178   2 u   14   64   37   92.632    0.623   0.799

timedatectl:

               Local time: Wed 2021-01-06 23:06:44 UTC
           Universal time: Wed 2021-01-06 23:06:44 UTC
                 RTC time: Wed 2021-01-06 23:06:44
                Time zone: Etc/UTC (UTC, +0000)
System clock synchronized: no
              NTP service: inactive
          RTC in local TZ: no

Any idea what could be wrong?

Kernel time error 0x41

I found some threads about this error, but solutions did not worked for me.

Tried sysctl -w kern.timecounter.hardware=ACPI-fast , but it did not help any. I reverted back to TSC.

My system is pfSense 2.6.0 on bare metal with Intel J4125 CPU. BIOS time and battery is also check’d.

Here is some logs and settings:

Anything to try?

That doesn’t look like a problem to me unless you’re actually seeing time inaccuracy?
It’s normal to see ntpd report that when it first starts.

If I restart it here for example:

Thanks Steve for reply.

Yes it have problem. It slowly drifting behind. This time it was almost two minutes behind and I had to stop NTP service and manually run ntpdate command. Logs are same before and after this manual correction.

Is it normal to see NTP request coming from pfBlockerNG virtual IP?

Did you see any other connections.. while you will see pool as a place holder, you should also see the actual servers your talking too with a reach of 377, etc.

I don’t run pfblocker vip, I just use it for aliases that I use in my own rules. But yeah that seems odd — is that vip actually natted to your public IP? If it tries and go to the public?

An intelligent man is sometimes forced to be drunk to spend time with his fools
If you get confused: Listen to the Music Play
Please don’t Chat/PM me for help, unless mod related
SG-4860 22.05 | Lab VMs CE 2.6, 2.7

There is no other than pools shown.

I have pfBlockerNG with quite stock settings. I have disabled automatic firewall policies and run it with these manually made floating rules:

Check Status > Interfaces. Are you seeing the WAN trying to use the pfBlocker VIP as it’s primary IP?
You might be hitting this: https://redmine.pfsense.org/issues/11545

There probably isn’t an outbound NAT rule for it so traffic to the actual NTP servers is failing.

Normally the pfBlocker VIP is on LAN so that should not be an issue.

Do you have specific interfaces set in NTP to listen on? It also uses those to send from and if you are only listening on LAN I could see that happening. Does it succeed if you deselect all the interfaces so it listens on all (the default)?

Can’t find 10.10.10.1 on any interface on interfaces status page.

I had spesified listen interfaces for NTP. I had understood that selected interfaces was only for NTP server listen interfaces.

Cleared and restarted server. Now it looks much more promising.

Is there a way to make sure everything is working?

Is there a way to make sure everything is working?

Yeah that looks better — it should fairly quickly show reach of 377, if so then its working correctly.

I really don’t see point to pointing to all of those pools. You should pick a pool for your region of the world and that should be more than enough..

An intelligent man is sometimes forced to be drunk to spend time with his fools
If you get confused: Listen to the Music Play
Please don’t Chat/PM me for help, unless mod related
SG-4860 22.05 | Lab VMs CE 2.6, 2.7

Thank you @stephenw10 and @johnpoz this looks like it is working now.

I assigned only one pool to NTP and now reach column shows 377 for four servers. So this is golden. Thanks again!

Источник

Arch Linux

You are not logged in.

#1 2020-07-01 23:49:56

System clock and Hardware Clock [SOLVED]

I was just hoping someone could help me work out If I’ve set the time correctly on my system. When I did the install I did

and after I installed the Desktop Environment I installed ntp and enabled ntpd.service.

I don’t know why there’s the kernel error saying clock unsynchronized.

when I set up reflector it seems to be going by UTC time

To be honest I don’t have a clue about all this time stuff and am totally confused. I just want to know if its configured correctly so I won’t have issues with it in the future. The time on my screen is correct for my local time.

Any help would be much appreciated.

Can someone help me correct the time if that is needed and explain how I can make sure it stays synchronized. I did think I had set the time correctly and that ntpd was going to keep it synchronized.

Anyone hopefully someone who doesn’t mind explaining things in a bit more detail without just directing me to the wiki. I do understand the intention of directing someone to the wiki. I have put it a serious amount of effort to just get a basic arch install and desktop environment setup. From scratch not long ago I started learning about Linux and only after installing arch have I asked a couple of questions.

I’m still stuck and quite confused with the whole issue of setting the time. Not only am I new to Arch but I am new to Linux too. I don’t want to give up on Arch now that I’ve come this far and spent so many weeks to get to where I am. I had read the appropriate wiki pages before I first posted here and then again after being directed back to them, but sadly the wiki isn’t written for absolute novices like me and I’ve struggled immensely trying to understand anything from it.

I have a few questions.

1) how does One clarify there system time and hardware clock is correct or wrong?

2) how to set up a time synchronisation service correctly so the time issue never needs to be a concern again?

3) If I was to do a fresh install using

and then after installing the desktop environment if I disabled

and then installed ntp and enabled

would/should all my time issues be resolved?

Last edited by Straight-path (2020-07-02 19:05:24)

Источник

ntpd -g does not sync the clock

From ntpd man page

If time is more than 1000s from the server time, ntpd assumes something must be terribly wrong and the only reliable action is for the operator to intervene and set the clock by hand. This causes ntpd to exit with a panic message to the system log. The -g option overrides this check and the clock will be set to the server time regardless of the chip time .

I have done small experiment to test -g option with ntpd. First I changed the system clock time to some old time with date command.

date -s 2021.06.15-19:10:21

After that I created small /etc/ntp.conf file with below information

After that I ran ntpd with below command

ntpd -g -n -4 -c /etc/ntp.conf &

Please note that my ntp.drift file was empty.

I see no change in the system time , infact ntp status shows that clock is not synchronized.

Can someone please help me. Did I missed any configuration or some other data.

Apart from this I have one small question

Does ntp clock need to be synchronised for ntp authentication? If ntp clock is not synchronised then in that case will ntp server authentication pass.

Below are the logs come when I start ntpd

Update:

@John I have done what ever you suggested but still I see same issue

Can you please check once. Did I missed anything?

Update

1 Answer 1

Normally after allowing NTP through any firewall, the first few packets are processed (increasing reach), a system peer is selected, and the initial step fixes the time. This system has reachable peers but rejected them all, which is odd.

After reviewing readvar output, rootdelay=0.000 doesn’t make sense. Remote over the internet, cannot be zero. Possibly why you got a flash word of distance threshold exceeded.

A theory from IRC is that NTP packets are being mangled by some middlebox. Which if true would be bad as it apparently broke NTP. ntpd applies dscp to its packets but ntpdate does not. Try dscp 0 in ntp.conf Ask around if differentiated services aware boxes are in the path. Take a packet capture (tcpdump) of NTP packets and look at it in Wireshark. See if it can dissect both any DSCP fields, wrapping NTP packets with NTP timestamps.

Also study ntpd in an entirely different network to contrast what it looks like when working. ntpd can run on anything, so maybe a VM on your workstation.

My original answer with a critique of the config follows.

Your reach recorded there is only 1. Wait 3 minutes after starting ntpd. Takes some time to get the first couple packets returned.

Regarding your configuration:

Consider adding iburst to your server and pool lines. Initial burst at startup of several packets with a short delay.

I am not convinced configuring minpool and maxpool is a good idea, neither is Red Hat. Too frequent for a sustained period and public NTP services may block you. ntpd is already good at dynamically selecting the pool interval.

Need more NTP servers, ideally 4 for detecting falsetickers with n+1 redundancy. Google public NTP has 4 frontends, which you can use with either the pool directive or multiple server lines. Feel free to add other NTP services as a second opinion, like 2.pool.ntp.org . (However you will have a messy leap second smear if your NTP servers don’t agree on that.)

No, public NTP services do not require NTP authentication. Authentication is not your problem, allowing large steps and tweaking the poll rate is.

My suggested change for the server line is more like this:

As to how you’re starting ntpd:

-g option is necessary to fix such a large offset of many hours, keep it. Allows infinite step once at start. Normally, large offset causes ntpd to panic. Use -g in any script that starts ntpd, so time will be fixed at boot.

I don’t see the purpose of no fork plus shell backgrounding. For starting in the background, I use the system’s service manager or init script. For example, a ntpd.service unit for Linux with systemd.

Delete -4 . Google is IPv6 capable.

If you wish to set the time once and exit, the -q option is useful. For interactive use when troubleshooting, normal operation is leaving ntpd running. Do not use ntpdate, which is obsolete.

Источник

Thread: NTP Issues

Thread Tools

Display

NTP Issues

Let me start out by saying I’m a noob at Linux as well as these forums.

I’m currently trying to update my NTP servers from 14.04 to 16.04. When I did the upgrade, it caused all havoc, and my network devices that rely heavily on NTP sent massive amounts of alerts. Thusly, I created new NTP servers from scratch, using 16.04. I’m using NTP Tool to troubleshoot client connectivity to the server. On the existing 14.04, I get no drops at 100 packets a second. On the new 16.04, I get 97-98 drops at 100 packets a second. The servers have synced to their NTP sources (ntp.org pools). This has really got me stumped, and any help will be greatly appreciated.

My syslogs for NTP.

The ntp.conf is mostly default, I commented out the Ubuntu servers and added the ntp.org pools (note below).

Here is the sync status of the servers.

Re: NTP Issues

Really sounds like a WAN connection issue, not specific to NTP.

BTW, more than 3 external servers is a waste, IMHO. If low msec correct time is that important, use a GPS device as source. Those numbers are huge! Use a local server and sync off it. That way, only 1 machine will be bouncing around and sending packets externally.

Oh . no need to sudo it. Over use of sudo can get you into trouble.

Where is the driftfile specified?
Do you restrict access to only the LAN systems?

I have 1 system that goes externally to get time from 3 pool servers (small network here). This box is the only time server for all the systems internally. They all point at it, including a few Windows systems. Just add the restrict and broadcast options to make an NTP server. Generally, perfectly accurate time really isn’t that important, so anything within a few minutes is close enough. Systems on the same network just need to closely have the same time. I’ve never seen any Unix machine off more than about 0.5 sec . though Windows systems seem to drift wildly for some reason. Too keep Windows times accurate, I had to force time checking every 15 minutes — anything longer than that, even 60 min, and the clock would be off 5+ minutes. Over a week, and it would be off almost an hour. The exact same system running Linux never has any time issues, so it is definitely a Windows problem. Plus this has happened to my Windows machines since the mid-90s. They just don’t keep time for some reason.

I agree with using non-Ubuntu time servers, though suspect they just point to the normal NTP pools. NTP is a good way for a distro to track running systems and the DNS redirect . well . those can be scary from a security standpoint.

Re: NTP Issues

I wish I could say it’s the WAN, but I’m testing internally and the company I am doing this for is an ISP.

Thanks for the note on «sudo», I get into the habit of sudoing everything, and I’m not about to open the root account. It’s a love/hate thing.

The drift file is the default location: /var/lib/ntp/ntp.drift. Driftfile output is 22.816.

I have not restricted access to any subnets that will be using it, yet. What I can’t understand, is that how is the 14.04 server allowing all these connections without issue, yet 16.04 will only answer the first 2-3 and drop the rest. I even disabled apparmor to see if that was the issue, but nope. Likewise with the firewall, it’s disabled. I assure you, minus the updates I listed in the original posts, both servers have the same ntp.conf setup.

We are running Calix devices that rely heavily on the NTP servers, and when I updated them a few weeks ago, they broke due to this. I personally don’t have experience with Calix stuff, but from what I could tell, it’s crazy. I could tell them to test use the Windows Server 2012 R2 hosts as the NTP servers (I forgot to mention, we are running the servers via Hyper-V). I’m half way tempted to just leave them on 14.04 for now. It’s just crazy that I can’t upgrade the server from 14.04 to 16.04 and have NTP not allow these devices to query for NTP.

Re: NTP Issues

VMs change everything with keeping time. There are usually guest additions that passthru the physical host’s time to all clients. But I don’t know **anything**, zero, nada, nichts about hyper-v. IME, Windows can’t keep time, so I’d never have it in my infrastructure where that was important.

Test if this is an issue on a physical 16.04 machine. I only have 1 16.04 install (zero servers and only a play chromebook) and it doesn’t show any time issues.
BTW — it appears that NTP changed between 16.04 and 14.04. Check the release notes. The «server» entry isn’t what is being used anymore. There is a «pool» entry now in the default configs that I don’t recall previously. We use ansible to maintain ntp.conf files across all the systems, so checking any others for differences won’t be any help.

A guess (and it is only that) is to look for hyper-v updates related to the 16.04 release.

Interesting and frustrating problem. At an ISP, time is critical, since you should provide that as a service to the clients. Setup 3 time servers internally and get 3 GPS-based time devices for those to sync. Then point all the others at those 3 machines. Did this a few decades ago on a network that wasn’t connected to the internet. Bet it is cheaper and easier now. I wouldn’t use any internet-based time pools at all.

Last edited by TheFu; September 8th, 2016 at 07:52 PM .

Re: NTP Issues

I have an old E-Machine laying around that I might be able to put Ubuntu on to test at home. If not, maybe I can get my Raspberry Pi to run the ARM version of 16.04. Or, I can turn my laptop into a NTP server (running desktop 16.04). I’ll test and report when I can get this setup. I should be good on the NTP server side though, I can see that I am connected to a Stratum 2 NTP server, which isn’t too shabby. There is no way that they will purchase an atomic clock, though it would be cool if they did.

Re: NTP Issues

$35 for a GPS bluetooth device doesn’t seem that expensive. That level of time accuracy should be plenty for everyone except crypto operators.
IMHO.
https://blog.logentries.com/2015/07/. ource-to-ntpd/

I have an old BT GPS receiver that needs to be repurposed for this use myself. Gotta keep learning.

Re: NTP Issues

That looks neat, but the issue isn’t getting Ubuntu to sync with an NTP server. The issue is getting other devices to connect to the Ubuntu server with NTP. However, I’ve always wanted to get a Raspberry Pi, a breadboard, and a GPS module to create an atomic clock.

Re: NTP Issues

I’m fairly certain that virtual-PC is the issue (or whatever they call it now).
Any luck changing the server to pool in the config files?

Re: NTP Issues

In case you are using iptables filtering on the ntp servers, did you double check them after the upgrade? In 16.04 the naming of the interfaces is changed so that might affect any services that are strictly linked to interface names. Maybe this was the issue?

Re: NTP Issues

I did change them, but I wasn’t any closer to getting the clients to connect.

In reference to the iptables, the current server, I’m unsure of (I would bet it’s not running, we use a Cisco ASA and just NAT the NTP server to an external IP). As for the new server, the one giving me the troubles, iptables is turned off (along with ufw).

Right now, I’m updating my Ubuntu desktop laptop (as I type. ) to set as a NTP server to see if using a physical box makes a difference.

UPDATE: I just got done testing, and it’s the same story. I’m going to revisit the firewall in the AM, just to make sure that it’s fully disabled. Also, I’ll create a 14.04 virtual machine to see if it does the same thing without configuring.

Last edited by hammons2012; September 9th, 2016 at 12:33 AM .

Источник

Kernel time error 0x41

I found some threads about this error, but solutions did not worked for me.

Tried sysctl -w kern.timecounter.hardware=ACPI-fast , but it did not help any. I reverted back to TSC.

My system is pfSense 2.6.0 on bare metal with Intel J4125 CPU. BIOS time and battery is also check’d.

Here is some logs and settings:

Anything to try?

That doesn’t look like a problem to me unless you’re actually seeing time inaccuracy?
It’s normal to see ntpd report that when it first starts.

If I restart it here for example:

Thanks Steve for reply.

Yes it have problem. It slowly drifting behind. This time it was almost two minutes behind and I had to stop NTP service and manually run ntpdate command. Logs are same before and after this manual correction.

Is it normal to see NTP request coming from pfBlockerNG virtual IP?

Did you see any other connections.. while you will see pool as a place holder, you should also see the actual servers your talking too with a reach of 377, etc.

I don’t run pfblocker vip, I just use it for aliases that I use in my own rules. But yeah that seems odd — is that vip actually natted to your public IP? If it tries and go to the public?

An intelligent man is sometimes forced to be drunk to spend time with his fools
If you get confused: Listen to the Music Play
Please don’t Chat/PM me for help, unless mod related
SG-4860 22.05 | Lab VMs CE 2.6, 2.7

There is no other than pools shown.

I have pfBlockerNG with quite stock settings. I have disabled automatic firewall policies and run it with these manually made floating rules:

Check Status > Interfaces. Are you seeing the WAN trying to use the pfBlocker VIP as it’s primary IP?
You might be hitting this: https://redmine.pfsense.org/issues/11545

There probably isn’t an outbound NAT rule for it so traffic to the actual NTP servers is failing.

Normally the pfBlocker VIP is on LAN so that should not be an issue.

Do you have specific interfaces set in NTP to listen on? It also uses those to send from and if you are only listening on LAN I could see that happening. Does it succeed if you deselect all the interfaces so it listens on all (the default)?

Can’t find 10.10.10.1 on any interface on interfaces status page.

I had spesified listen interfaces for NTP. I had understood that selected interfaces was only for NTP server listen interfaces.

Cleared and restarted server. Now it looks much more promising.

Is there a way to make sure everything is working?

Is there a way to make sure everything is working?

Yeah that looks better — it should fairly quickly show reach of 377, if so then its working correctly.

I really don’t see point to pointing to all of those pools. You should pick a pool for your region of the world and that should be more than enough..

An intelligent man is sometimes forced to be drunk to spend time with his fools
If you get confused: Listen to the Music Play
Please don’t Chat/PM me for help, unless mod related
SG-4860 22.05 | Lab VMs CE 2.6, 2.7

Thank you @stephenw10 and @johnpoz this looks like it is working now.

I assigned only one pool to NTP and now reach column shows 377 for four servers. So this is golden. Thanks again!

Источник

Thread: NTP Issues

Thread Tools

Display

NTP Issues

Let me start out by saying I’m a noob at Linux as well as these forums.

I’m currently trying to update my NTP servers from 14.04 to 16.04. When I did the upgrade, it caused all havoc, and my network devices that rely heavily on NTP sent massive amounts of alerts. Thusly, I created new NTP servers from scratch, using 16.04. I’m using NTP Tool to troubleshoot client connectivity to the server. On the existing 14.04, I get no drops at 100 packets a second. On the new 16.04, I get 97-98 drops at 100 packets a second. The servers have synced to their NTP sources (ntp.org pools). This has really got me stumped, and any help will be greatly appreciated.

My syslogs for NTP.

The ntp.conf is mostly default, I commented out the Ubuntu servers and added the ntp.org pools (note below).

Here is the sync status of the servers.

Re: NTP Issues

Really sounds like a WAN connection issue, not specific to NTP.

BTW, more than 3 external servers is a waste, IMHO. If low msec correct time is that important, use a GPS device as source. Those numbers are huge! Use a local server and sync off it. That way, only 1 machine will be bouncing around and sending packets externally.

Oh . no need to sudo it. Over use of sudo can get you into trouble.

Where is the driftfile specified?
Do you restrict access to only the LAN systems?

I have 1 system that goes externally to get time from 3 pool servers (small network here). This box is the only time server for all the systems internally. They all point at it, including a few Windows systems. Just add the restrict and broadcast options to make an NTP server. Generally, perfectly accurate time really isn’t that important, so anything within a few minutes is close enough. Systems on the same network just need to closely have the same time. I’ve never seen any Unix machine off more than about 0.5 sec . though Windows systems seem to drift wildly for some reason. Too keep Windows times accurate, I had to force time checking every 15 minutes — anything longer than that, even 60 min, and the clock would be off 5+ minutes. Over a week, and it would be off almost an hour. The exact same system running Linux never has any time issues, so it is definitely a Windows problem. Plus this has happened to my Windows machines since the mid-90s. They just don’t keep time for some reason.

I agree with using non-Ubuntu time servers, though suspect they just point to the normal NTP pools. NTP is a good way for a distro to track running systems and the DNS redirect . well . those can be scary from a security standpoint.

Re: NTP Issues

I wish I could say it’s the WAN, but I’m testing internally and the company I am doing this for is an ISP.

Thanks for the note on «sudo», I get into the habit of sudoing everything, and I’m not about to open the root account. It’s a love/hate thing.

The drift file is the default location: /var/lib/ntp/ntp.drift. Driftfile output is 22.816.

I have not restricted access to any subnets that will be using it, yet. What I can’t understand, is that how is the 14.04 server allowing all these connections without issue, yet 16.04 will only answer the first 2-3 and drop the rest. I even disabled apparmor to see if that was the issue, but nope. Likewise with the firewall, it’s disabled. I assure you, minus the updates I listed in the original posts, both servers have the same ntp.conf setup.

We are running Calix devices that rely heavily on the NTP servers, and when I updated them a few weeks ago, they broke due to this. I personally don’t have experience with Calix stuff, but from what I could tell, it’s crazy. I could tell them to test use the Windows Server 2012 R2 hosts as the NTP servers (I forgot to mention, we are running the servers via Hyper-V). I’m half way tempted to just leave them on 14.04 for now. It’s just crazy that I can’t upgrade the server from 14.04 to 16.04 and have NTP not allow these devices to query for NTP.

Re: NTP Issues

VMs change everything with keeping time. There are usually guest additions that passthru the physical host’s time to all clients. But I don’t know **anything**, zero, nada, nichts about hyper-v. IME, Windows can’t keep time, so I’d never have it in my infrastructure where that was important.

Test if this is an issue on a physical 16.04 machine. I only have 1 16.04 install (zero servers and only a play chromebook) and it doesn’t show any time issues.
BTW — it appears that NTP changed between 16.04 and 14.04. Check the release notes. The «server» entry isn’t what is being used anymore. There is a «pool» entry now in the default configs that I don’t recall previously. We use ansible to maintain ntp.conf files across all the systems, so checking any others for differences won’t be any help.

A guess (and it is only that) is to look for hyper-v updates related to the 16.04 release.

Interesting and frustrating problem. At an ISP, time is critical, since you should provide that as a service to the clients. Setup 3 time servers internally and get 3 GPS-based time devices for those to sync. Then point all the others at those 3 machines. Did this a few decades ago on a network that wasn’t connected to the internet. Bet it is cheaper and easier now. I wouldn’t use any internet-based time pools at all.

Last edited by TheFu; September 8th, 2016 at 07:52 PM .

Re: NTP Issues

I have an old E-Machine laying around that I might be able to put Ubuntu on to test at home. If not, maybe I can get my Raspberry Pi to run the ARM version of 16.04. Or, I can turn my laptop into a NTP server (running desktop 16.04). I’ll test and report when I can get this setup. I should be good on the NTP server side though, I can see that I am connected to a Stratum 2 NTP server, which isn’t too shabby. There is no way that they will purchase an atomic clock, though it would be cool if they did.

Re: NTP Issues

$35 for a GPS bluetooth device doesn’t seem that expensive. That level of time accuracy should be plenty for everyone except crypto operators.
IMHO.
https://blog.logentries.com/2015/07/. ource-to-ntpd/

I have an old BT GPS receiver that needs to be repurposed for this use myself. Gotta keep learning.

Re: NTP Issues

That looks neat, but the issue isn’t getting Ubuntu to sync with an NTP server. The issue is getting other devices to connect to the Ubuntu server with NTP. However, I’ve always wanted to get a Raspberry Pi, a breadboard, and a GPS module to create an atomic clock.

Re: NTP Issues

I’m fairly certain that virtual-PC is the issue (or whatever they call it now).
Any luck changing the server to pool in the config files?

Re: NTP Issues

In case you are using iptables filtering on the ntp servers, did you double check them after the upgrade? In 16.04 the naming of the interfaces is changed so that might affect any services that are strictly linked to interface names. Maybe this was the issue?

Re: NTP Issues

I did change them, but I wasn’t any closer to getting the clients to connect.

In reference to the iptables, the current server, I’m unsure of (I would bet it’s not running, we use a Cisco ASA and just NAT the NTP server to an external IP). As for the new server, the one giving me the troubles, iptables is turned off (along with ufw).

Right now, I’m updating my Ubuntu desktop laptop (as I type. ) to set as a NTP server to see if using a physical box makes a difference.

UPDATE: I just got done testing, and it’s the same story. I’m going to revisit the firewall in the AM, just to make sure that it’s fully disabled. Also, I’ll create a 14.04 virtual machine to see if it does the same thing without configuring.

Last edited by hammons2012; September 9th, 2016 at 12:33 AM .

Источник

Kernel time error 0x41

I cannot set the RPi to sync its time over the local network. I have one RPi that is setup as
NTP server. This is OK and it is sending its time to the Hubitat Elevation when it’s being requested.
The server part is then OK. I’m trying to get the others RPI to sync their time from the
same local server. There is no internet on this network.

The way I understand this, the RPi comes pre-installed with timesyncd service. So I modified the file:

where i did set the ip address of my NTP server. (By the way, the NTP server is using the NTP service)

I tried this, but didn’t work. I then uncomment the following 3 lines where they set the poll interval etc.
Did not work either (even after rebooting).

I also tried to stop and start the service with

sudo systemctl stop systemd-timesyncd
sudo systemctl start systemd-timesyncd

did not work either. Strange enough, the status shows

NTP service: active
System clock synchronised: no

but the time is never synchronised.

So I gave up with timesyncd and tried with NTP. I first disabled the timesyncd

sudo systemctl stop systemd-timesyncd
sudo systemctl disable systemd-timesyncd

and installed NTP

sudo apt-get install ntp

I edited the configuration file

sudo nano /etc/ntp.conf

in that file, I added a line for my server

and comment out the 4 poll lines

rebooted and requested some outputs:

remote refid st t when poll reach delay offset jitter
================================================================================
192.168.2.34 .XFAC. 16 u 26 64 0 0.000 0.000 0.000

192.168.2.34 : IP of my server (OK)
.XFAC. : Source of time. (I guess it is still undefined)
16 : Stratum 16 means untrusted time source
u : Unicast
26 : When: Number of second since last response. This is changing all the time, so I assume it is talking to it
64 : Poll: Interval in seconds between time request
0 : Reach: indicates success/failure to reach the source

the other values mean nothing since it does not reach the source.

Notes: This is a new fresh install Debian and all the updates done.. The ping to the server is responding
properly. I also rebooted and retried the ntpq command and now the «when» value is staying «-«

Re: Cannot get NTP client to sync with local server

Re: Cannot get NTP client to sync with local server

It would make sense if others couldn’t talk to the server as well. But HE does get the time form the ntp server.

Everything is the default; even the router is restored to the default just to be sure.

Re: Cannot get NTP client to sync with local server

I also disabled all firewall, and nothing more.

The log shows also this:

jui 12 00:03:16 raspberrypi systemd[1]: Started Network Time Service.
jui 12 00:03:16 raspberrypi ntpd[1070]: proto: precision = 0.759 usec (-20)
jui 12 00:03:16 raspberrypi ntpd[1070]: leapsecond file (‘/usr/share/zoneinfo/leap-seconds.list’): good hash signature
jui 12 00:03:16 raspberrypi ntpd[1070]: leapsecond file (‘/usr/share/zoneinfo/leap-seconds.list’): loaded, expire=2020-12-28T00:00:00Z last=2017-01-01T00:00:00Z ofs=37
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen and drop on 0 v6wildcard [::]:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen and drop on 1 v4wildcard 0.0.0.0:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen normally on 2 lo 127.0.0.1:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen normally on 3 wlan0 192.168.2.27:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen normally on 4 lo [::1]:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen normally on 5 wlan0 [fe80::dfc1:4a0:c4ad:cf5f%3]:123
jui 12 00:03:16 raspberrypi ntpd[1070]: Listening on routing socket on fd #22 for interface updates
jui 12 00:03:16 raspberrypi ntpd[1070]: Listen for broadcasts to 192.168.2.255 on interface #3 wlan0
jui 12 00:03:16 raspberrypi ntpd[1070]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
jui 12 00:03:16 raspberrypi ntpd[1070]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized

Re: Cannot get NTP client to sync with local server

Re: Cannot get NTP client to sync with local server

Re: Cannot get NTP client to sync with local server

Re: Cannot get NTP client to sync with local server

Omg. i ran to the rpi to try adjusting the clock 10 minutes off the server. But no . Same problem. I rebooted and same problem.

But i didn’t understand where to put the -g option.

I am wondering also why in the log i dont see the ip of the server it is supposed to poll the time from. Somehow i thought it would make sense to write that there. Unless im not configured properly.

Re: Cannot get NTP client to sync with local server

Do you have the correct timezone set in raspi-config?

Reset the clock by running

Re: Cannot get NTP client to sync with local server

Ok, you want me to remove the ntp package and go back to timedatectl?

I beleive they dont work well together these two services.

Re: Cannot get NTP client to sync with local server

Re: Cannot get NTP client to sync with local server

Samething. I see that the datetimectl command did effectively changed the time but after a reboot, it failed again.

Compared to the log i posted earlier, there that line that bugs me a lot related to no interface available for broadcast.

Re: Cannot get NTP client to sync with local server

Samething. I see that the datetimectl command did effectively changed the time but after a reboot, it failed again.

Compared to the log i posted earlier, there that line that bugs me a lot related to no interface available for broadcast.

I am not sure this is a failure, ntp takes a little while to synchronize. It’s been a while, but I want to say it was 5-10 samplings, which depending on your polling settings could take 5 minutes or hours. Usually most systems would do an ntpdate before starting ntpd to sync the clock up, and ntpd just makes the adjustments for drift. I believe you have to install the ntpdate package, and tweak /etc/default/ntpdate if you want to sync the clock at start up, then ntpd would start up and keep it in sync.

This is why I am wondering why you abandoned timesyncd, it does the forced sync which your are going to need due the absence of a hardware clock and it keeps the clock in sync. I would say at this point, maybe way an hour or 2 and check your syslog and ntpq to see if the status changes.

Re: Cannot get NTP client to sync with local server

grrrr. I wont have any hair left after this.

Thanks a lot for helping me with this. I’m desperated.

I did wait and wait, but no time update. But by looking at the ntpq command, it sounds like it’s never reaching the server. I also installed the ntpdate but no luck. When i type

sudo ntpdate 192.168.2.34

I see it trying for a few seconds and then:

«no server suitable for synchronization found»

I tried to find if there was a way to see the requests coming in the server but didn’t find how. If I could see this, I could compare with the Hubitat requests. And I could probably see if the server is indeed answering, but only the client is ignoring it. Can’t find anything.

As far as timesyncd, I only abandoned it because I could not make it work either. I really dont mind which one to use.

The strange thing is that this is bran new install, completly redone from scratch, and installed the updates. That’s all I’ve done to it. It is surprising that it can’t work right off the box. What am I doing wrong?

The router is an old Cisco E2000. But Hubitat uses it too and it works. That’s why I concentrate only on the client.

Re: Cannot get NTP client to sync with local server

One thing I noticed is the ntp server at 192.168.2.34 was reporting stratum 16, now I don’t know if that is real or not, I want to say it starts with 16 until it can actually talk to the ntp server to probe it’s stratum. But if it is 16, clients will not sync.

I think it was suggested earlier, but this is sounding a bit like a firewall issue. Perhaps your Cisco router has ACLs in place to restrict access? Do the public NTP servers not work?

Re: Cannot get NTP client to sync with local server

The firewall in the Cisco is turned off. Beside, if Hubitat can go through, the RPi should.

And yes, 16 means untrusted time source. But who decide it’s 16? I dont recall having set that anywhere.

To try the public server, I’ll have to connect this network to internet. I’ll do that. it might help finding that out.

Re: Cannot get NTP client to sync with local server

public servers works. Not the one you gave me, but this one:

I get errors in the log but it does work anyway.

In the list, that debian server stayed as a stratum 16, but I can see a list of about 12 ip addresses below with st being 1, 2 or 3

I dont know what to think anymore.

Re: Cannot get NTP client to sync with local server

The firewall in the Cisco is turned off. Beside, if Hubitat can go through, the RPi should.

And yes, 16 means untrusted time source. But who decide it’s 16? I dont recall having set that anywhere.

To try the public server, I’ll have to connect this network to internet. I’ll do that. it might help finding that out.

Well the way stratum works is let’s say server A is using a reference clock, so he is stratum 1, server B syncs from A so he is stratum 2 server C syncs from B so he is stratum 3, and you sync to C so you are stratum 4. I am fairly sure the reference clock driver will end up forcing your server to 1, I have never done a sync to a «reference clock» but it appears be 127.127.x.y IPs where x is the reference driver — so i am guessing it declares itself as 0 so the NTPd server connecting to the reference clock will advertise itself as 1. http://doc.ntp.org/4.1.2/refclock.htm

If you do not sync to another server then you are usually stratum 16. So if your Cisco router is not syncing a time server it will be stratum 16, if that is by design and you are just using a system/hardware clock then you can fudge the stratum to 1 but knowing the time your serving will not be accurate and will drift over time (and all of the downstream clients will drift). I don’t know about Cisco routers, but for ntpd you can sync to your system clock (i think the server is 127.127.1.1, it’s usually in the comments) and then you use the fudge command to fudge the stratum.

Whatever this device is may not have a proper ntp implementation, but ntpd will refuse to sync to a stratum 16 server.

Re: Cannot get NTP client to sync with local server

wow. this is getting out of hand

The only thing I want is to get all the computers at the same date and time (plus or minus 10 minutes). Since I had to have a ntp server to set the clock of the Hubitat hub, I thought to use the same server to update the other clients. All the docs I read about this seems to be very simple. But it’s not.

I think I can write a SYNC program in python that will do the job. I’m on this case simce 6 am today. Writing a python will only take me half an hour. But I was trying to use the CANE software

Thanks for your help.

Re: Cannot get NTP client to sync with local server

Yeah it’s usually pretty simple. In my case systemd-timesyncd runs on all my Linux boxes and they all sync to a public NTP server of some sort, generally whatever the OS has configured out of the box.

Years ago I used to run ntpd for all my local devices, but honestly couldn’t be bothered anymore.

Re: Cannot get NTP client to sync with local server

In case of unsurmountable firewall problem you can try using htpdate instead of ntpdate.
It uses headers in the HTTP response from servers to set a date by bumping the local clock.
Not as accurate as NTP but good enough.

(Not all http servers include the needed headers, if the date doesn’t adjust, try hitting other ones.)

Re: Cannot get NTP client to sync with local server

Ok guys. Thanks a lot for helping.

I wasn’t able to give up without a last try. So a remove ntp and tried again using the timesyncd. Surprising enough, i was able to find out that the client is indeed receiving something from the server. Because if In the server’s ntp.conf, i set the line

restrict 192.168.2.1 mask 255.255.255.255 nomodify notrap

I cant remember the exact syntax but there was a mention in the comment that we must restrict this only if cryptographically authenticated. Anyway, if i set this line, the timesyncd was reporting a timeout trying to get the time. Recommenting it then i was not getting this time out. So, somehow, the client is receiving some kind of answer back from the server. But i could not get any farther. I tried and tried lots of different setup.

So, i disabled timesyncd and reinstalled ntp. Doing the same change in the sever i could see the POLL value of the ntpd command incrementing when the restriction was not active, and not incrementing when the restriction was ON. This is also means that it is receiving something. but what?

Strange that there are no log to report what is really going on. Anyhow, i tried so many different things. i ended up writing a python program on the server to broadcast the time, and in the clients, i get it and update its time. This is at least working. Not exactly the way i prefer but it is working.

If someone read this one day and have a clue what was happening here, let me know.

Источник

Hey,

I have an error with my NTP daemon. When I check its status for the first time with

systemctl status ntpd.service

here is the output:

Oct 26 17:14:42 myhost ntpd[377]: unable to create socket on enp2s0 (5) for fe80::493f:8ce3:2992:16b5%2#123
Oct 26 17:14:42 myhost ntpd[377]: failed to init interface for address fe80::493f:8ce3:2992:16b5%2
Oct 26 17:14:42 myhost ntpd[377]: new interface(s) found: waking up resolver
Oct 26 17:14:44 myhost ntpd[377]: bind(24) AF_INET6 xxxx:xxxx:xx:xxxx:xxxx:xxxx:xxxx:xxxx#123 flags 0x11 failed: Cannot assign re>
Oct 26 17:14:44 myhost ntpd[377]: unable to create socket on enp2s0 (6) for xxxx:xxxx:xx:xxxx:xxxx:xxxx:xxxx:xxxx#123
Oct 26 17:14:44 myhost ntpd[377]: failed to init interface for address xxxx:xxxx:xx:xxxx:xxxx:xxxx:xxxx:xxxx
Oct 26 17:14:44 myhost ntpd[377]: Listen normally on 7 enp2s0 [fe80::493f:8ce3:2992:16b5%2]:123
Oct 26 17:14:44 myhost ntpd[377]: new interface(s) found: waking up resolver
Oct 26 17:14:46 myhost ntpd[377]: Listen normally on 8 enp2s0 [xxxx:xxxx:xx:xxxx:xxxx:xxxx:xxxx:xxxx]:123
Oct 26 17:14:46 myhost ntpd[377]: new interface(s) found: waking up resolver

It seems to be an IPv6 error.

Then, when I restart the service, and that I check its status again, I get another one:

Oct 26 17:18:13 myhost ntpd[1178]: Listen and drop on 0 v6wildcard [::]:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen and drop on 1 v4wildcard 0.0.0.0:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen normally on 2 lo 127.0.0.1:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen normally on 3 enp2s0 192.168.1.25:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen normally on 4 lo [::1]:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen normally on 5 enp2s0 [xxxx:xxxx:xx:xxxx:xxxx:xxxx:xxxx:xxxx]:123
Oct 26 17:18:13 myhost ntpd[1178]: Listen normally on 6 enp2s0 [fe80::493f:8ce3:2992:16b5%2]:123
Oct 26 17:18:13 myhost ntpd[1178]: Listening on routing socket on fd #23 for interface updates
Oct 26 17:18:13 myhost ntpd[1178]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
Oct 26 17:18:13 myhost ntpd[1178]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized

This time it is a kernel time error.

My configuration file `/etc/ntp.conf`:

# Please consider joining the pool:
#
#     http://www.pool.ntp.org/join.html
#
# For additional information see:
# - https://wiki.archlinux.org/index.php/Network_Time_Protocol_daemon
# - http://support.ntp.org/bin/view/Support/GettingStarted
# - the ntp.conf man page

# Associate to Arch's NTP pool
server 0.fr.pool.ntp.org iburst
server 1.fr.pool.ntp.org iburst
server 2.fr.pool.ntp.org iburst
server 3.fr.pool.ntp.org iburst

# By default, the server allows:
# - all queries from the local host
# - only time queries from remote hosts, protected by rate limiting and kod
restrict default kod limited nomodify nopeer noquery notrap
restrict 127.0.0.1
restrict ::1

# Location of drift file
driftfile /var/lib/ntp/ntp.drift

The output of `ntpq -pn`

     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
+185.123.84.51   200.98.196.212   2 u   30   64  377   15.617   -0.422  20.934
+80.67.184.40    85.199.214.102   2 u   29   64  377    4.370    0.859   9.489
*2001:bc8:313f:: 145.238.203.14   2 u   26   64  377    4.182    1.177   8.150
+178.32.220.152  138.96.64.10     2 u   25   64  377    7.363    3.703   8.125

Thank you in advance for your help.

Last edited by Wmog (2018-10-26 15:37:04)