Every day at 22pm all links down

RouterOS General
1

Every day all active links down :frowning: I listed all scheduled tasks, but can’t found any in this time.
May be this is ISP joke, but why all links down? Any ideas?

nov/24 22:00:11 interface,info wan link down
nov/24 22:00:11 interface,info box link down
nov/24 22:00:11 interface,info rodos link down
nov/24 22:00:11 interface,info oops link down
nov/24 22:00:11 dhcp,info dhcp-client on wan lost IP address xx.37.184.187 - lease stopped locally
nov/24 22:00:14 interface,info box link up (speed 1G, full duplex)
nov/24 22:00:14 interface,info rodos link up (speed 10M, half duplex)
nov/24 22:00:14 e-mail,error Error sending e-mail <Main DHCP alert!>: DNS resolve failed
nov/24 22:00:15 interface,info wan link up (speed 1G, full duplex)
nov/24 22:00:15 interface,info oops link up (speed 1G, full duplex)
nov/24 22:00:15 dhcp,info dhcp-client on wan got IP address xx.37.184.187
nov/24 22:00:17 dhcp,info dhcp deassigned 192.168.88.2 from xx:xx:CA:05:08:17
nov/24 22:00:17 dhcp,info dhcps assigned 192.168.88.2 to xx:xx:CA:05:08:17
nov/24 22:00:18 dhcp,info dhcp deassigned 192.168.88.70 from xx:xx:2B:36:82:92
nov/24 22:00:18 dhcp,info dhcps assigned 192.168.88.70 to xx:xx:xx:36:82:92

2

Sorry,

       routerboard: yes
             model: RouterBOARD 3011UiAS
  factory-firmware: 3.41
  current-firmware: 6.45.7
  upgrade-firmware: 6.45.7

Alert is script on DHCP client…

3

As I inverstigated, it’s all interfaces on one (first) switch.
Can wan link disconnection propagated to all other links?

4

Normally not, one port going down doesn’t affect others. But maybe there is some broadcast storm and it causes the switch to go completely down for a while - it’s just a speculation.

5

May be this is watchdog? Can it down links, not reboot?

6

Definitely not, watchdog reboots the whole machine when it stops giving it treats (because it is dead). There is no watchdog for a single switch chip unless they’ve silently added it recently, and if they did, you’d see it in the log.

Is it feasible for you to migrate the cables from ether1-5 to ether6-10 one-by-one to see whether the issue is related to one of them, or rather the traffic coming from that cable?