Can;t access router after reboot until switch reset

Started by kermitxyz, December 06, 2024, 01:33:30 AM

Previous topic - Next topic


I have just changed my main router to OPNSense on a new x86 device.  After rebooting the opensense router, I can't get any response (ping/GUI) from the OPNSense device unless I reset the switch on my network core.  Once the switch reboots, all is well (can access the router and WAN interface comes up)

My first thought was some MAC / ARP issue, but having power-cycled the whole core the problem persists.

I know this is odd - does anyone have any thoughts?

December 07, 2024, 12:10:01 AM #1 Last Edit: December 07, 2024, 12:20:13 AM by kermitxyz
A re-install of OPNSense fixed the issue.

Just for completeness, the only explanation of which I can think is that the router and switch were struggling to auto-negotiate on the port and a reboot of the switch somehow fixed that.  I recall a similar incident nearly twenty years ago when trying to get some PCs to connect to a Cisco switch.  We just couldn't work out why certain machine didn't work until enabling portfast on the ports in question.

If anyone can think of any other possible reasons I remain curious....

December 12, 2024, 11:59:16 PM #2 Last Edit: December 13, 2024, 12:11:48 AM by kermitxyz
Changing from the on-board NIC (Fujitsu S920) to another network port in the router seems to have resolved the issue.  The router and switch indeed couldn't auto-negotiate.  Perhaps a driver issue?

Quote from: kermitxyz on December 12, 2024, 11:59:16 PMChanging from the on-board NIC (Fujitsu S920) to another network port in the router seems to have resolved the issue.  The router and switch indeed couldn't auto-negotiate.  Perhaps a driver issue?
No help for you, but here's a similar weird one..
My Asustor NAS kept on dropping the connection (2.5 Gb/s) to the switch (cheap Chinese one). Rebooting either the switch of the NAS would bring it back up. It seemed to fail when I was hammering the NAS (backing up to the cloud at 2.5 Gb/s) so I suspected that something was overheating somewhere. I ordered a new switch. In the meantime one of the disks in the NAS failed with the "click of death", so I swapped it out. No NIC problems ever since so I have a spare switch!
Maybe the failing disk was drawing more current or sending spurious noise back over the SATA or power lines, who knows?

The Fujitsu S920 reportedly uses a Realtek NIC, so why wonder (see #6)?
Intel N100, 4 x I226-V, 16 GByte, 256 GByte NVME, ZTE F6005

1100 down / 440 up, Bufferbloat A+