0

I have a problem with some of my PIs losing Ethernet connection. This happens between every few days to a month. After the Ethernet connection is lost, it does not come back up. Besides the Ethernet connection being lost, the PIs continue running normally.

It only effects Raspberry Pi 3B (without +) models. They all have good, high current power supplies. I am not using WiFi. The Raspberries with the problem are all revision a02082, 3 Model B, 1 GB, Overvolt : No, Released : Q1 2016, Notes : (Mfg by Sony). With Raspbian GNU/Linux 10 (buster) and latest Raspberry OS kernel Linux 5.10.60-v7+ #1449 SMP Wed Aug 25 15:00:01 BST 2021 armv7l GNU/Linux, but it also affects earlier kernels.

It happens too frequently to be a freak accident. The problem affects 3 different boards at different locations. The cables and switches look fine. This is a especially problematic if they are running headless somewhere as a server and I need to manually powercycle them.

See this sample log of around the time of one of those connection losses:

Sep 16 14:59:26 my-rp dhcpcd[434]: eth0: carrier lost
Sep 16 14:59:26 my-rp dhcpcd[434]: eth0: deleting address fe80::61cc:1701:<REMOVED MAC>:ca73
Sep 16 14:59:27 my-rp kernel: [223066.725645] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabledSep 16 14:59:28 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k eth0
Sep 16 14:59:28 my-rp dhcpcd[434]: eth0: removing interface
Sep 16 14:59:28 my-rp kernel: [223066.924906] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 14:59:28 my-rp kernel: [223067.550363] SMSC LAN8700 usb-001:003:01: attached PHY driver [SMSC LAN8700] (mii_bus:phy_addr=usb-001:003:01, irq=POLL)
> Sep 16 14:59:28 my-rp kernel: [223067.550752] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 14:59:28 my-rp kernel: [223067.561248] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 14:59:28 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: eth0  error: No such file or directory
Sep 16 14:59:31 my-rp kernel: [223070.408930] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled
Sep 16 15:00:01 my-rp CRON[8659]: (smmsp) CMD (test -x /etc/init.d/sendmail && test -x /usr/share/sendmail/sendmail && test -x /usr/lib/sm.bin/sendmail && /usr/share/sendmail/sendmail cron-msp)
Sep 16 15:00:01 my-rp sm-msp-queue[8679]: My unqualified host name (my-rp) unknown; sleeping for retry
Sep 16 15:01:01 my-rp sm-msp-queue[8679]: unable to qualify my own domain name (my-rp) -- using short name
Sep 16 15:01:39 my-rp kernel: [223198.390821] smsc95xx 1-1.1:1.0 eth0: Link is Up - 100Mbps/Full - flow control off
Sep 16 15:01:39 my-rp kernel: [223198.390893] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Sep 16 15:03:06 my-rp wicd[428]: sending commands to master dhcpcd process
Sep 16 15:03:06 my-rp dhcpcd[8766]: sending commands to master dhcpcd process
Sep 16 15:03:06 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k wlan0
Sep 16 15:03:07 my-rp kernel: [223286.149956] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled
Sep 16 15:03:07 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: wlan0  error: No such file or directory
Sep 16 15:03:07 my-rp wicd[428]: sending commands to master dhcpcd process
Sep 16 15:03:07 my-rp dhcpcd[8789]: sending commands to master dhcpcd process
Sep 16 15:03:07 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k eth0
Sep 16 15:03:07 my-rp kernel: [223286.337556] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 15:03:07 my-rp kernel: [223286.338316] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:07 my-rp kernel: [223286.455796] SMSC LAN8700 usb-001:003:01: attached PHY driver [SMSC LAN8700] (mii_bus:phy_addr=usb-001:003:01, irq=POLL)
Sep 16 15:03:07 my-rp kernel: [223286.456274] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:07 my-rp kernel: [223286.466706] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 15:03:07 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: eth0  error: No such file or directory
Sep 16 15:03:07 my-rp kernel: [223286.591683] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:07 my-rp wicd[428]: sending commands to master dhcpcd process
Sep 16 15:03:07 my-rp dhcpcd[8815]: sending commands to master dhcpcd process
Sep 16 15:03:07 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k eth0
Sep 16 15:03:07 my-rp kernel: [223286.723781] SMSC LAN8700 usb-001:003:01: attached PHY driver [SMSC LAN8700] (mii_bus:phy_addr=usb-001:003:01, irq=POLL)
Sep 16 15:03:07 my-rp kernel: [223286.724180] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:07 my-rp kernel: [223286.734646] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 15:03:07 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: eth0  error: No such file or directory
Sep 16 15:03:10 my-rp dhcpcd[8825]: sending commands to master dhcpcd process
Sep 16 15:03:10 my-rp dhcpcd[434]: control command: /sbin/dhcpcd --noipv4ll eth0
Sep 16 15:03:12 my-rp kernel: [223290.839397] smsc95xx 1-1.1:1.0 eth0: Link is Up - 100Mbps/Full - flow control off
Sep 16 15:03:12 my-rp kernel: [223290.839469] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: carrier acquired
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: IAID eb:eb:7c:a3
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: adding address fe80::61cc:1701:<REMOVED MAC>:ca73
Sep 16 15:03:12 my-rp avahi-daemon[396]: Joining mDNS multicast group on interface eth0.IPv6 with address fe80::61cc:1701:<REMOVED MAC>:ca73.
Sep 16 15:03:12 my-rp avahi-daemon[396]: New relevant interface eth0.IPv6 for mDNS.
Sep 16 15:03:12 my-rp avahi-daemon[396]: Registering new address record for fe80::61cc:1701:<REMOVED MAC>:ca73 on eth0.*.
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: soliciting a DHCP lease
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: offered 192.168.193.9 from 192.168.193.10
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: probing address 192.168.193.9/24
Sep 16 15:03:12 my-rp dhcpcd[434]: eth0: soliciting an IPv6 router
Sep 16 15:03:14 my-rp wicd[428]: sending commands to master dhcpcd process
Sep 16 15:03:14 my-rp dhcpcd[8854]: sending commands to master dhcpcd process
Sep 16 15:03:14 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k wlan0
Sep 16 15:03:15 my-rp kernel: [223294.053802] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled
Sep 16 15:03:15 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: wlan0  error: No such file or directory
Sep 16 15:03:15 my-rp wicd[428]: sending commands to master dhcpcd process
Sep 16 15:03:15 my-rp dhcpcd[8877]: sending commands to master dhcpcd process
Sep 16 15:03:15 my-rp dhcpcd[434]: control command: /sbin/dhcpcd -k eth0
Sep 16 15:03:15 my-rp dhcpcd[434]: eth0: removing interface
Sep 16 15:03:15 my-rp dhcpcd[434]: eth0: deleting address fe80::61cc:1701:<REMOVED MAC>:ca73
Sep 16 15:03:15 my-rp avahi-daemon[396]: Withdrawing address record for fe80::61cc:1701:<REMOVED MAC>:ca73 on eth0.
Sep 16 15:03:15 my-rp avahi-daemon[396]: Leaving mDNS multicast group on interface eth0.IPv6 with address fe80::61cc:1701:<REMOVED MAC>:ca73.
Sep 16 15:03:15 my-rp avahi-daemon[396]: Interface eth0.IPv6 no longer relevant for mDNS.
Sep 16 15:03:15 my-rp kernel: [223294.248643] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 15:03:15 my-rp kernel: [223294.253131] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:15 my-rp kernel: [223294.371834] SMSC LAN8700 usb-001:003:01: attached PHY driver [SMSC LAN8700] (mii_bus:phy_addr=usb-001:003:01, irq=POLL)
Sep 16 15:03:15 my-rp kernel: [223294.372236] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Sep 16 15:03:15 my-rp kernel: [223294.382680] smsc95xx 1-1.1:1.0 eth0: Link is Down
Sep 16 15:03:15 my-rp wicd[428]: Failed to connect to non-global ctrl_ifname: eth0  error: No such file or directory
Sep 16 15:03:17 my-rp kernel: [223296.438546] smsc95xx 1-1.1:1.0 eth0: Link is Up - 100Mbps/Full - flow control off
Sep 16 15:03:17 my-rp kernel: [223296.438590] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Sep 16 15:17:01 my-rp CRON[9250]: (root) CMD (   cd / && run-parts --report /etc/cron.hourly)
Sep 16 15:20:01 my-rp CRON[9334]: (smmsp) CMD (test -x /etc/init.d/sendmail && test -x /usr/share/sendmail/sendmail && test -x /usr/lib/sm.bin/sendmail && /usr/share/sendmail/sendmail cron-msp)
Sep 16 15:20:01 my-rp sm-msp-queue[9353]: My unqualified host name (my-rp) unknown; sleeping for retry

Removed unrelated loglines and part of the MAC address. I am not sure if the issue starts with eth0: carrier lost or with smsc95xx 1-1.1:1.0 eth0: Link is Down. In any case the connectivity never recovers without manual intervention.

Could this be a hardware or kernel error? What can I do to fix this issue?

7
  • You've mis-configured something. Tell us what networking configuration changes you've made, and we'll try to help you sort this.
    – Seamus
    Commented Sep 28, 2021 at 17:11
  • what happens when you schedule a reboot every 24 hours? There might be some cleanup items that you have set to be done on boot instead of regularly.
    – Abel
    Commented Sep 29, 2021 at 1:04
  • @Seamus Thanks! No network configuration changes made, the Pis get their IP from DHCP.
    – user643011
    Commented Sep 29, 2021 at 9:51
  • @Abel If I reboot the Pi then the network starts working again. But until the scheduled reboot the network remains disconnected. What cleanup items should I check? Do the log error messages give you any idea what could be the problem?
    – user643011
    Commented Sep 29, 2021 at 9:54
  • 1
    Since I haven't heard anything any more I am going to replace the Pi3s with Pi4s soon. The Pi4s are rock solid compared to the Pi3s and it's not worth the hassle keeping the Pi3s running with those connectivity losses vs the few bucks it costs to buy Pi4s. If you have a suggestion where else to report this Ethernet NIC issue (maybe it's a kernel bug or a hardware bug) or I can provide some more logs or run tests to help the community then now would be the right time to say so.
    – user643011
    Commented Oct 21, 2021 at 16:00

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.