Network Interface Randomly Drops

Support
1

AlmaLinix 9.7. Server has 2 physical 1G network cards, one with 2 ports (E1), the other with 4 (E2). At present, E1 connects to a switch on one subnet and has done so for years. E2 directly plugs into a Synology NAS for backups only using a different subnet for only those two items and this is new.

For some reason, the E2 ethernet will occasionally go down. It can successfully be restarted either through the desktop or ifconfig eno12399 up.

Can’t figure out why this happens, and not even sure where to look. Closest I get is /var/log/messages, but have no idea what it’s telling me:

Feb 21 13:49:07 landru kernel: tg3 0000:8a:00.0 eno12399: Link is up at 1000 Mbps, full duplexFeb 21 13:49:07 landru kernel: tg3 0000:8a:00.0 eno12399: Flow control is on for TX and on for RXFeb 21 13:49:07 landru kernel: tg3 0000:8a:00.0 eno12399: EEE is disabledFeb 21 13:49:07 landru vmnet-natd[92992]: RTM_NEWLINK: name:eno12399 index:4 flags:0x00011043Feb 21 13:49:07 landru NetworkManager[2561]:   [1771706947.9658] device (eno12399): carrier: link connectedFeb 21 13:49:07 landru vmnetBridge[92973]: RTM_NEWLINK: name:eno12399 index:4 flags:0x00011043Feb 21 13:49:07 landru NetworkManager[2561]:   [1771706947.9663] device (eno12399): state change: unavailable → disconnected (reason ‘carrier-changed’, managed-type: ‘full’)Feb 21 13:49:07 landru vmnetBridge[92973]: Adding interface eno12399 index:4

Any suggestions on where to begin?

2

Hello. That log (“Link is up at 1000 Mbps”) only indicates it “recovered” and does not state why it dropped.

Please post the kernel logs and NetworkManager logs for the 2-3 minutes before and after the time it dropped.

Kernel
journalctl -k --since “YYYY-MM-DD HH:MM” --until “YYYY-MM-DD HH:MM” | egrep -i ‘tg3|eno12399|watchdog|reset|timeout|link is’

NetworkManager
journalctl -u NetworkManager --since “YYYY-MM-DD HH:MM” --until “YYYY-MM-DD HH:MM”