ntpMerlin ntpMerlin - NTP Daemon for AsusWRT Merlin

[joe68000]

Thanks! I'll give it a try.

 

[L&LD]

Also, shouldn't 'iburst' be on each line too?

 

[Kingp1n]

Great points! With ntpMerlin, I am at ~51.54 u offset. I would try the built-in NTP server but the graphs are too nice and I wouldn't know how accurate they would be, comparatively!

Would having both enabled mess anything up? Having the script installed with the built-in NTP server enabled?

 

[L&LD]

Would having both enabled mess anything up? Having the script installed with the built-in NTP server enabled?

I'm sure it would mess things up.

But turning off one or the other, then rebooting the router (I would do it twice, waiting 15 minutes between reboots) would probably fix it.

 

[L&LD]

All Jack Yaz scripts used (please see my signature below) working with no issues on 384.12 Alpha 1.

https://www.snbforums.com/threads/384-12_alpha-builds-testing-all-variants.56639/#post-491571

 

[joe68000]

It might be that you have too many servers in your conf?
Pool will manage the number to be about eight.
Comment out your eight pool servers and use:

pool north-america.pool.ntp.org iburst
server time-a-g.nist.gov iburst
server time-b-g.nist.gov iburst

The pool will include the two NIST servers which are close to you along with six servers from the pool.

No matter how I configured ntpd (using pools, specifying servers, or both) it would not stay synced with any peer. I finally found a thread that was discussing this somewhere else, and the solution proposed was to set:

tos maxdist 16

So I tried it and it solved the problem. My ntpd now syncs with a peer and stays synced.

 

[Swistheater]

What is the leap file option that is commented out inside the ntp.conf?

 

[Swistheater]

No matter how I configured ntpd (using pools, specifying servers, or both) it would not stay synced with any peer. I finally found a thread that was discussing this somewhere else, and the solution proposed was to set:

tos maxdist 16

So I tried it and it solved the problem. My ntpd now syncs with a peer and stays synced.

I concur to some degree with this setup but i believe you should use
pool time.nist.gov iburst because this will allow the ntp to use a wider range of servers and will automatically disable it if it becomes unresponsive or if the time goes wrong where if you use the server option this will cause more issues

also you could try <--- if you have any home servers you want fall back to if everything else fails.
#point to localhost if access is lost to NTP servers/pools
fudge IPofHost stratum #

 

[vdemarco]

What is the leap file option that is commented out inside the ntp.conf?

You can download the file, (I did) the last time there was a leap second event was in 2017.
There isn't one in the near future (at least that's my read of the file)

 

[joe68000]

I concur to some degree with this setup but i believe you should use
pool time.nist.gov iburst because this will allow the ntp to use a wider range of servers and will automatically disable it if it becomes unresponsive or if the time goes wrong where if you use the server option this will cause more issues

also you could try <--- if you have any home servers you want fall back to if everything else fails.
#point to localhost if access is lost to NTP servers/pools
fudge IPofHost stratum #

I reconfigured to use this pool and removed all of the servers. So far, so good. Thanks!

 

[EmeraldDeer]

I reconfigured to use this pool and removed all of the servers. So far, so good. Thanks!

You have an unusual network problem which the "tos maxdist 16" config option works around.

While changing servers around may or may not be a good idea, I would be surprised if it had much impact.

I speculate that there are frequent but intermittent delays. I would not want to be a gamer on your network. Perhaps connmon configured to your first hop ISP router could inform.

 

[EmeraldDeer]

You can download the file, (I did) the last time there was a leap second event was in 2017.
There isn't one in the near future (at least that's my read of the file)

• I have not noticed any problem in other's NTP daemons without a leap file
• Having it does set the TAI offset to the current correct value of 37
• As you said, the next one is not planned yet
• If one is planned, the leap file allows the daemon itself and connected clients to know it is coming
• Leap seconds are needed because the rotation of the earth is slowing
• When there is a leap second the clock goes:
  • 23:59:59
  • 23:59:60
  • 00:00:00
• http://www.leapsecond.com/java/gpsclock.htm shows local, UTC, GPS and TAI
• If leap second happens without the leap file, then the clock will instantaneously be one second ahead. At best, the clock will be slowed until synchronization.

 

[joe68000]

You have an unusual network problem which the "tos maxdist 16" config option works around.

While changing servers around may or may not be a good idea, I would be surprised if it had much impact.

I speculate that there are frequent but intermittent delays. I would not want to be a gamer on your network. Perhaps connmon configured to your first hop ISP router could inform.

Interesting... no complaints from the gamers on this network. I installed connmon and set it to the first hop to my ISP (FIOS):

connmon: Test results - Ping 11.675 ms - Jitter - 4.930 ms - Line Quality 100.000 %

I'll keep my eye on it.

 

[joe68000]

Is there an easy way to test if all network ntp traffic is being directed to ntpMerlin (option 2 is enabled)?

 

[Swistheater]

Is there an easy way to test if all network ntp traffic is being directed to ntpMerlin (option 2 is enabled)?

Do a ntp update on a client like a laptop or desktop while observing the changes on the connections page under syslogs tab. You should see traffic from that devices ip pointing at the routers ip via port 123.

 

[joe68000]

Thanks! Of course, I forgot about simply checking the open connections. I can clearly see all of the devices on my network with udp connections to the router on 123.

 

[doczenith1]

Hmmmmm I see this after I clicked the update now button.

The IP address of the Win 10 computer I clicked the update now button on is 192.168.1.4 and I have the time server set to router.asus.com. .174 is one of my arlo cameras so looks like my ntp traffic is being directed to the router. I can't seem to get my Win 10 computer's IP address to show up in the list...until...I switch the time server back to time.windows.com and then:

So looks like things are working as designed.

 

[Swistheater]

note if your clock is way off from the routers just wait until it successfully gets time you may have to click the update time option a couple of times before getting a successful sync

 

[EmeraldDeer]

As @Swistheater has mentioned in another thread, the NAT redirect gets lost sometimes. Here is toggling it off:

Choose an option:    2

iptables: No chain/target/match by that name.
NTP Redirect has been disabled

Press enter to continue...

Here is toggling it back on:

Choose an option:    2

NTP Redirect has been enabled

Press enter to continue...

 

[Jack Yaz]

Weird, it's created via nat-start so should persist