Time keeping in the event of a power failure

RouterOS General
1

I don’t have my PoE injector connected to a UPS and we tend to have frequent power failures.

Already while testing my new setup, I have noted that after setting the time and shutting down the CPE to unplug the PoE, the time set in WinBox is lost. Thus the next time I open up the Log I see Jan/1970 timestamps again! (Bridge not yet fully in production 24/7 use)

I have studied the RouterBoard circuit diagram and there appears to be no battery backup like on a PC motherboard (BIOS settings).

Can I synchronise the RouterBoard time with an atomic clock of sorts? i.e. point the RouterBoard to a time keeping service?

Edit Typo’s fixed

2

the only thing you can do is make it synchronize from a GPS or from a NTP server (time.nist.gov, for example, is a free server you can use)

3

To update for any South African users:

Time Servers in ZA can be found here but note you can only link to a “Stratum-2 Server”. I found ntp.is.za to have public access and my RouterBoard is being updated every 64s.

[admin@MikroTik] /system ntp client> print
        enabled: yes
           mode: broadcast
    primary-ntp: 196.4.160.4
  secondary-ntp: 0.0.0.0

Not sure if “broadcast” mode stresses the hell out of the RouterBoard? Things became very slow all of a sudden - I mean really SLOW. :confused: (Hope it’s this and not my ROS upgrade from 3.6 → 3.11)

Updated:

[admin@MikroTik] /system ntp client> print
             enabled: yes
                mode: unicast
         primary-ntp: 196.4.160.4
       secondary-ntp: 196.2.45.66
       poll-interval: 2m8s
       active-server: 196.2.45.66
    last-update-from: 196.2.45.66
  last-update-before: 43s470ms
     last-adjustment: -12ms147us

Polling interval seems to adjust itself - now at 256s.

4

I’m in Jozi and these work perfectly;

mode: unicast
primary-ntp: 196.4.160.4 (IS)
secondary-ntp: 196.31.48.139 (UUNET)

5

I’m in the sticks and these also work fine:-


Unicast
P:146.64.24.58
S:196.26.5.8

Hilton - we’ve seen erratic behaviour on that IS one the last 2 weeks or so… just an idle observation.
We have a LOT of BSS / Clients and I don’t want them “hunting” an address that doesn’t work.
But - I’m known to be be "overengineeredpedantic.. :slight_smile:

T

6

when in doubt, just visit;

http://www.time.za.net/

On principle, I won’t use MWEB’s NTP server.

7

Sup guys, thanks for the replies. My NTP has gone dead and I cannot trace my steps. I also deleted all my config backups under System > Files in favour of keeping only the latest backup. When it was working, my ADSL router had DHCP server turned on and my whole network was getting IP’s from the pool. Then I started to get small IP errors and reserved IP’s went haywire. I eventually disabled DHCP Server and put my whole network on static IP’s. Then I came to notice that NTP was no longer working.

I can see that the RouterBoards are trying to poll the Primary NTP and Secondary NTP every 16s but get no connection.

Can anyone tell me what NTP is reliant on? Also, if my ADSL router is using its own NTP, can I set each RouterBoard to get the time from the ADSL router?