[opensuse] Intel 82574L NIC Problem
Hi all, I am having a major problem on one of my systems with a network controller. The system has the Intel DQ77MK motherboard. This board has dual NICs. The first NIC, an Intel 82574LM works fine and is called eno1 The second NIC, an intel 82574L is the problem. The second NIC is called rename3, or sometimes rename2 / rename6 or rename7 etc.. It keeps changing. Whenever I try to apply any configuration to this NIC it changes name again, therefore I have no network connection. So I change the name in the config and the NIC changes name again. I have never seen a NIC called rename before. What is this trickery?!?!? The system is running ubuntu 18.04. Now I have probably confused you by posting on the OpenSUSE list, but the fact is I have posted on the Ubuntu forum and it has been 4 days and I have had no reply what so ever. It is still Linux after all, so hopefully one of you might be able to assist? Ubuntu 18.04 uses netplan to manage the network configuration. I intend to set up an bond in balance-rr mode as I have done on another system without issue. How can I go about diagnosing this? This is the first time I have every had a NIC not work in any linux distro (apart from broadcom wifi drives cough cough..) so I have literally no in-depth knowledge of networking in linux. so basically... HELP! lol First things I checked: Output of lspci: paul@s34:~$ lspci | grep net 00:19.0 Ethernet controller: Intel Corporation 82579LM Gigabit Network Connection (rev 04) 03:00.0 Ethernet controller: Intel Corporation 82574L Gigabit Network Connection Output of ifconfig: paul@s34:~$ ifconfig -a eno1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.0.164 netmask 255.255.255.0 broadcast 192.168.0.255 inet6 fe80::7e05:7ff:fe0f:11c4 prefixlen 64 scopeid 0x20<link> ether 7c:05:07:0f:11:c4 txqueuelen 1000 (Ethernet) RX packets 1049 bytes 106842 (106.8 KB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 315 bytes 44842 (44.8 KB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 device interrupt 20 memory 0xf7f00000-f7f20000 lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10<host> loop txqueuelen 1000 (Local Loopback) RX packets 178 bytes 12392 (12.3 KB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 178 bytes 12392 (12.3 KB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 rename9: flags=4098<BROADCAST,MULTICAST> mtu 1500 ether 7c:05:07:0f:11:c5 txqueuelen 1000 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 device interrupt 18 memory 0xf7e00000-f7e20000 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
22.10.2018 20:29, Paul Groves пишет:
Hi all,
I am having a major problem on one of my systems with a network controller.
The system has the Intel DQ77MK motherboard. This board has dual NICs. The first NIC, an Intel 82574LM works fine and is called eno1 The second NIC, an intel 82574L is the problem.
The second NIC is called rename3, or sometimes rename2 / rename6 or rename7 etc.. It keeps changing.
Whenever I try to apply any configuration to this NIC it changes name again, therefore I have no network connection. So I change the name in the config and the NIC changes name again.
I have never seen a NIC called rename before. What is this trickery?!?!?
This is Debian specific patch. Number is kernel interface index. "rename%u" is supposed to be temporary until interface got renamed to final name.
The system is running ubuntu 18.04. Now I have probably confused you by posting on the OpenSUSE list, but the fact is I have posted on the Ubuntu forum and it has been 4 days and I have had no reply what so ever. It is still Linux after all, so hopefully one of you might be able to assist?
Streetlight effect ...
Ubuntu 18.04 uses netplan to manage the network configuration. I intend to set up an bond in balance-rr mode as I have done on another system without issue.
How can I go about diagnosing this?
I usually start with checking logs. udev fails to rename interface; logs should contain error message with error number (or error string) which may give some starting point. Booting with "debug" on kernel command line (or enabling it specifically for udevd) could be helpful too.
This is the first time I have every had a NIC not work in any linux distro (apart from broadcom wifi drives cough cough..) so I have literally no in-depth knowledge of networking in linux.
so basically... HELP! lol
First things I checked:
Output of lspci: paul@s34:~$ lspci | grep net 00:19.0 Ethernet controller: Intel Corporation 82579LM Gigabit Network Connection (rev 04) 03:00.0 Ethernet controller: Intel Corporation 82574L Gigabit Network Connection
Output of ifconfig: paul@s34:~$ ifconfig -a eno1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.0.164 netmask 255.255.255.0 broadcast 192.168.0.255 inet6 fe80::7e05:7ff:fe0f:11c4 prefixlen 64 scopeid 0x20<link> ether 7c:05:07:0f:11:c4 txqueuelen 1000 (Ethernet) RX packets 1049 bytes 106842 (106.8 KB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 315 bytes 44842 (44.8 KB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 device interrupt 20 memory 0xf7f00000-f7f20000
lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10<host> loop txqueuelen 1000 (Local Loopback) RX packets 178 bytes 12392 (12.3 KB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 178 bytes 12392 (12.3 KB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
rename9: flags=4098<BROADCAST,MULTICAST> mtu 1500 ether 7c:05:07:0f:11:c5 txqueuelen 1000 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 device interrupt 18 memory 0xf7e00000-f7e20000
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/10/2018 19:25, Andrei Borzenkov wrote:
This is Debian specific patch. Number is kernel interface index. "rename%u" is supposed to be temporary until interface got renamed to final name.
Well this at least explains why I have never seen this. Normally use SUSE so never come across this before. Output from udev rules: paul@s34:~$ cat /lib/udev/rules.d/80-net-setup-link.rules # do not edit this file, it will be overwritten on update SUBSYSTEM!="net", GOTO="net_setup_link_end" IMPORT{builtin}="path_id" ACTION!="add", GOTO="net_setup_link_end" IMPORT{builtin}="net_setup_link" NAME=="", ENV{ID_NET_NAME}!="", NAME="$env{ID_NET_NAME}" LABEL="net_setup_link_end"
Streetlight effect ...
Almost... As I said, already posted to Ubuntu forum. Is there an Ubuntu list like this one? I am please to say that the OpenSUSE list has never failed me so far :)
I usually start with checking logs. udev fails to rename interface; logs should contain error message with error number (or error string) which may give some starting point. Booting with "debug" on kernel command line (or enabling it specifically for udevd) could be helpful too.
OK, so I assume you are saying that udev is what is configuring my NICs at boot? How do I check the udev log? I have tried the following. Where do I check the udev log? When I disable the secondary NIC in the BIOS everything is normal. When it is enabled booting takes about 3 mins longer. I had a look at dmesg. See the output relevant to intel drive below. The weird thing is, in the BIOS ASPM OS control is on. But dmesg still says it does not have access. Strange... So now I have disabled ASPM in the BIOS. Those two lines have gone from dmesg but the rest of the messages are still the same. So it is not that. Also, you can see where the NICs are renamed from eth0 and eth1 to eno0 and rename3. So it doesn't appear to have failed to rename them at that stage. Output from Dmesg: paul@s34:~$ dmesg | grep e1000e [ 0.958274] e1000e: Intel(R) PRO/1000 Network Driver - 3.2.6-k [ 0.958362] e1000e: Copyright(c) 1999 - 2015 Intel Corporation. [ 0.958595] e1000e 0000:00:19.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode [ 1.055224] e1000e 0000:00:19.0 0000:00:19.0 (uninitialized): registered PHC clock [ 1.143128] e1000e 0000:00:19.0 eth0: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c4 [ 1.143189] e1000e 0000:00:19.0 eth0: Intel(R) PRO/1000 Network Connection [ 1.143261] e1000e 0000:00:19.0 eth0: MAC: 10, PHY: 11, PBA No: FFFFFF-0FF [ 1.143337] e1000e 0000:03:00.0: Disabling ASPM L1 [ 1.143379] e1000e 0000:03:00.0: can't disable ASPM; OS doesn't have ASPM control [ 1.143629] e1000e 0000:03:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode [ 1.198213] e1000e 0000:03:00.0 0000:03:00.0 (uninitialized): registered PHC clock [ 1.251840] e1000e 0000:03:00.0 eth1: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c5 [ 1.251844] e1000e 0000:03:00.0 eth1: Intel(R) PRO/1000 Network Connection [ 1.251930] e1000e 0000:03:00.0 eth1: MAC: 3, PHY: 8, PBA No: FFFFFF-0FF [ 1.252554] e1000e 0000:00:19.0 eno1: renamed from eth0 [ 1.280245] e1000e 0000:03:00.0 rename3: renamed from eth1 [ 107.974069] e1000e: eno1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx [ 107.974072] e1000e 0000:00:19.0 eno1: 10/100 speed: disabling TSO -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
22.10.2018 22:24, Paul Groves пишет:
On 22/10/2018 19:25, Andrei Borzenkov wrote:
This is Debian specific patch. Number is kernel interface index. "rename%u" is supposed to be temporary until interface got renamed to final name.
Well this at least explains why I have never seen this. Normally use SUSE so never come across this before.
Output from udev rules: paul@s34:~$ cat /lib/udev/rules.d/80-net-setup-link.rules # do not edit this file, it will be overwritten on update
SUBSYSTEM!="net", GOTO="net_setup_link_end"
IMPORT{builtin}="path_id"
ACTION!="add", GOTO="net_setup_link_end"
IMPORT{builtin}="net_setup_link"
NAME=="", ENV{ID_NET_NAME}!="", NAME="$env{ID_NET_NAME}"
LABEL="net_setup_link_end"
Streetlight effect ...
Almost... As I said, already posted to Ubuntu forum. Is there an Ubuntu list like this one? I am please to say that the OpenSUSE list has never failed me so far :)
I usually start with checking logs. udev fails to rename interface; logs should contain error message with error number (or error string) which may give some starting point. Booting with "debug" on kernel command line (or enabling it specifically for udevd) could be helpful too.
OK, so I assume you are saying that udev is what is configuring my NICs at boot?
Where did I say it? udev is responsible for creating device aliases and renaming network interfaces. In your case renaming of network interface fails for whatever reason.
How do I check the udev log?
On every systemd-based system the first place to check is journal. If the question is "how do I recognize udev messages" - "journalctl -u systemd-udevd.service" filters only messages that come from processes spawned by udev. Or you simply upload "journalctl -b" output to let others to look at it.
I have tried the following. Where do I check the udev log? When I disable the secondary NIC in the BIOS everything is normal. When it is enabled booting takes about 3 mins longer.
If you see "renameX" interface name, the initial attempt to name interface failed and udev will retry for 90 seconds. This may account for at least half of this time.
I had a look at dmesg.
Only the very early messages before journald is started may appear in dmesg. dmesg is kernel and we speak about user space.
See the output relevant to intel drive below.
My usual reaction - if you knew what is relevant, you would not need to ask in the first place. Boot with "debug printk.devkmsg=on" and upload "journalctl -b" output.
The weird thing is, in the BIOS ASPM OS control is on. But dmesg still says it does not have access. Strange...
So now I have disabled ASPM in the BIOS. Those two lines have gone from dmesg but the rest of the messages are still the same. So it is not that.
Also, you can see where the NICs are renamed from eth0 and eth1 to eno0 and rename3. So it doesn't appear to have failed to rename them at that stage.
Output from Dmesg: paul@s34:~$ dmesg | grep e1000e [ 0.958274] e1000e: Intel(R) PRO/1000 Network Driver - 3.2.6-k [ 0.958362] e1000e: Copyright(c) 1999 - 2015 Intel Corporation. [ 0.958595] e1000e 0000:00:19.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode [ 1.055224] e1000e 0000:00:19.0 0000:00:19.0 (uninitialized): registered PHC clock [ 1.143128] e1000e 0000:00:19.0 eth0: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c4 [ 1.143189] e1000e 0000:00:19.0 eth0: Intel(R) PRO/1000 Network Connection [ 1.143261] e1000e 0000:00:19.0 eth0: MAC: 10, PHY: 11, PBA No: FFFFFF-0FF [ 1.143337] e1000e 0000:03:00.0: Disabling ASPM L1 [ 1.143379] e1000e 0000:03:00.0: can't disable ASPM; OS doesn't have ASPM control [ 1.143629] e1000e 0000:03:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode [ 1.198213] e1000e 0000:03:00.0 0000:03:00.0 (uninitialized): registered PHC clock [ 1.251840] e1000e 0000:03:00.0 eth1: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c5 [ 1.251844] e1000e 0000:03:00.0 eth1: Intel(R) PRO/1000 Network Connection [ 1.251930] e1000e 0000:03:00.0 eth1: MAC: 3, PHY: 8, PBA No: FFFFFF-0FF [ 1.252554] e1000e 0000:00:19.0 eno1: renamed from eth0 [ 1.280245] e1000e 0000:03:00.0 rename3: renamed from eth1 [ 107.974069] e1000e: eno1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx [ 107.974072] e1000e 0000:00:19.0 eno1: 10/100 speed: disabling TSO
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 05:40, Andrei Borzenkov wrote:
On every systemd-based system the first place to check is journal. If the question is "how do I recognize udev messages" - "journalctl -u systemd-udevd.service" filters only messages that come from processes spawned by udev.
OK here is the output from that command: -- Reboot -- Oct 23 15:13:53 media systemd[1]: Starting udev Kernel Device Manager... Oct 23 15:13:54 media systemd[1]: Started udev Kernel Device Manager. Oct 23 15:13:56 media systemd-udevd[487]: link_config: autonegotiation is unset or enabled, the speed and duplex are not writable. Oct 23 15:13:56 media systemd-udevd[479]: link_config: autonegotiation is unset or enabled, the speed and duplex are not writable. Oct 23 15:13:56 media systemd-udevd[484]: link_config: autonegotiation is unset or enabled, the speed and duplex are not writable.
Or you simply upload "journalctl -b" output to let others to look at it.
I has a scan through journalctl -b Here are all the things I think are relevant. I can always upload the full output if required. The hardware id of the secondary NIC is 03:00.0 It is initially named eth1 then renamed to rename3 paul@s34:~$ journalctl -b | grep -E 'e10001e|eth0|eth1|eno1|rename3|03:00.0|network' Oct 23 15:13:52 media kernel: pci 0000:03:00.0: [8086:10d3] type 00 class 0x020000 Oct 23 15:13:52 media kernel: pci 0000:03:00.0: reg 0x10: [mem 0xf7e00000-0xf7e1ffff] Oct 23 15:13:52 media kernel: pci 0000:03:00.0: reg 0x18: [io 0xd000-0xd01f] Oct 23 15:13:52 media kernel: pci 0000:03:00.0: reg 0x1c: [mem 0xf7e20000-0xf7e23fff] Oct 23 15:13:52 media kernel: pci 0000:03:00.0: PME# supported from D0 D3hot D3cold Oct 23 15:13:52 media kernel: e1000e 0000:00:19.0 eth0: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c4 Oct 23 15:13:52 media kernel: e1000e 0000:00:19.0 eth0: Intel(R) PRO/1000 Network Connection Oct 23 15:13:52 media kernel: e1000e 0000:00:19.0 eth0: MAC: 10, PHY: 11, PBA No: FFFFFF-0FF Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0 0000:03:00.0 (uninitialized): registered PHC clock Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0 eth1: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c5 Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0 eth1: Intel(R) PRO/1000 Network Connection Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0 eth1: MAC: 3, PHY: 8, PBA No: FFFFFF-0FF Oct 23 15:13:52 media kernel: e1000e 0000:00:19.0 eno1: renamed from eth0 Oct 23 15:13:52 media kernel: e1000e 0000:03:00.0 rename3: renamed from eth1 Oct 23 15:14:05 media systemd-networkd[787]: Enumeration completed Oct 23 15:14:05 media systemd-networkd[787]: lo: Link is not managed by us Oct 23 15:14:05 media systemd-networkd[787]: rename3: Link is not managed by us Oct 23 15:14:05 media systemd-networkd[787]: eno1: IPv6 successfully enabled Oct 23 15:14:05 media kernel: IPv6: ADDRCONF(NETDEV_UP): eno1: link is not ready Oct 23 15:14:07 media systemd-networkd[787]: eno1: Gained carrier Oct 23 15:14:07 media kernel: e1000e: eno1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx Oct 23 15:14:07 media kernel: e1000e 0000:00:19.0 eno1: 10/100 speed: disabling TSO Oct 23 15:14:07 media kernel: IPv6: ADDRCONF(NETDEV_CHANGE): eno1: link becomes ready Oct 23 15:14:07 media systemd-networkd[787]: eno1: DHCPv4 address 192.168.0.164/24 via 192.168.0.1 Oct 23 15:14:07 media systemd-networkd[787]: Not connected to system bus, not setting hostname. Oct 23 15:14:08 media systemd-networkd[787]: eno1: Gained IPv6LL Oct 23 15:14:08 media systemd-networkd[787]: eno1: Configured Oct 23 15:14:08 media systemd-networkd-wait-online[811]: managing: eno1 Oct 23 15:14:08 media systemd-networkd-wait-online[811]: ignoring: lo Oct 23 15:14:09 media dbus-daemon[917]: [system] Activating via systemd: service name='org.freedesktop.hostname1' unit='dbus-org.freedesktop.hostname1.service' requested by ':1.0' (uid=100 pid=787 comm="/lib/systemd/systemd-networkd " label="unconfined") Oct 23 15:14:10 media systemd[1]: Starting Dispatcher daemon for systemd-networkd... Oct 23 15:14:11 media networkd-dispatcher[1036]: No valid path found for iwconfig Oct 23 15:14:12 media systemd[1]: Started Dispatcher daemon for systemd-networkd. Oct 23 15:14:30 media systemd[2180]: Listening on GnuPG network certificate management daemon.
If you see "renameX" interface name, the initial attempt to name interface failed and udev will retry for 90 seconds. This may account for at least half of this time.
Maybe it is trying multiple times? Just a guess.
Only the very early messages before journald is started may appear in dmesg. dmesg is kernel and we speak about user space.
OK
My usual reaction - if you knew what is relevant, you would not need to ask in the first place.
Boot with "debug printk.devkmsg=on" and upload "journalctl -b" output.
Where do I add the option "debug printk.devkmsg=on" ? If I remember correctly you can do this in GRUB at boot time but that was 2010 when I last looked into that. Is this right? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Paul Groves wrote:
Boot with "debug printk.devkmsg=on" and upload "journalctl -b" output.
Where do I add the option "debug printk.devkmsg=on" ?
If I remember correctly you can do this in GRUB at boot time but that was 2010 when I last looked into that. Is this right?
Yep, those are kernel arguments - may be added permanently or on boot-up. -- Per Jessen, Zürich (13.2°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 17:23, Per Jessen wrote:
Paul Groves wrote:
Boot with "debug printk.devkmsg=on" and upload "journalctl -b" output.
Where do I add the option "debug printk.devkmsg=on" ?
If I remember correctly you can do this in GRUB at boot time but that was 2010 when I last looked into that. Is this right?
Yep, those are kernel arguments - may be added permanently or on boot-up.
OK so I rebooted and tried to hold shift but that didn't take me into grub. So I rebooted again and smashed the up and down arrows. That did take me into grub. So then I selected ubuntu and pressed e The I added: debug printk.devkmsg=on on the end Now here is my output from journalctl -b https://nextcloud.paulgroves.uk/index.php/s/ypcx7zHiTPcN69G Yb2oMzjE -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
23.10.2018 20:35, Paul Groves пишет:
On 23/10/2018 17:23, Per Jessen wrote:
Paul Groves wrote:
Boot with "debug printk.devkmsg=on" and upload "journalctl -b" output.
Where do I add the option "debug printk.devkmsg=on" ?
If I remember correctly you can do this in GRUB at boot time but that was 2010 when I last looked into that. Is this right?
Yep, those are kernel arguments - may be added permanently or on boot-up.
OK so I rebooted and tried to hold shift but that didn't take me into grub. So I rebooted again and smashed the up and down arrows. That did take me into grub.
So then I selected ubuntu and pressed e
The I added: debug printk.devkmsg=on on the end
Now here is my output from journalctl -b
https://nextcloud.paulgroves.uk/index.php/s/ypcx7zHiTPcN69G
Yb2oMzjE
Oct 23 17:25:23 media kernel: e1000e 0000:00:19.0 eno1: renamed from eth0 ... Oct 23 17:25:23 media systemd-udevd[158]: error changing net interface name 'eth1' to 'eno1': File exists Oct 23 17:25:23 media systemd-udevd[158]: could not rename interface '3' from 'eth1' to 'eno1': File exists Check /etc/udev/rules.d (70-persistent-net.rules is the primary suspect) for any file that contains eno1. Also check *.link files in various directories (/lib/systemd/network/, /etc/systemd/network; see "man systemd.link" for full list). Apparently both your interfaces are configured to have name "eno1". One possible reason could be firmware bug that assigns identical index to several interfaces. Please upload full output of "udevadm info --export-db". And please do it on https://susepaste.org/, not on some commercial site. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 18:48, Andrei Borzenkov wrote:
Check /etc/udev/rules.d (70-persistent-net.rules is the primary suspect)
Does not exist paul@media:~$ ls /etc/udev/rules.d/ 70-snap.core.rules
for any file that contains eno1. Also check *.link files in various directories (/lib/systemd/network/, /etc/systemd/network; see "man systemd.link" for full list).Apparently both your interfaces are configured to have name "eno1".
Nothing in these files: paul@media:~$ cat /lib/systemd/network/* | grep eno1 paul@media:~$ ls /lib/systemd/network/ 80-container-host0.network 80-container-ve.network 80-container-vz.network 99-default.link Empty: paul@media:~$ ls /etc/systemd/network/ That would explain the problem if they are configured to have the same name. What should I be looking for in man systemd.link?
And please do it onhttps://susepaste.org/, not on some commercial site. That is my own server actually. I can see it from here :)
I never heard of susepaste.org I am going to confuse the matter more now. I just rebooted the system and now the two NICs are reversed. The primary NIC is now called rename2 and is no longer working. The second NIC which has not worked until no is now called eno1 and is working fine. :? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
23.10.2018 21:05, Paul Groves пишет:
On 23/10/2018 18:48, Andrei Borzenkov wrote:
Check /etc/udev/rules.d (70-persistent-net.rules is the primary suspect)
Does not exist paul@media:~$ ls /etc/udev/rules.d/ 70-snap.core.rules
for any file that contains eno1. Also check *.link files in various directories (/lib/systemd/network/, /etc/systemd/network; see "man systemd.link" for full list).Apparently both your interfaces are configured to have name "eno1".
Nothing in these files: paul@media:~$ cat /lib/systemd/network/* | grep eno1 paul@media:~$ ls /lib/systemd/network/ 80-container-host0.network 80-container-ve.network 80-container-vz.network 99-default.link
Empty: paul@media:~$ ls /etc/systemd/network/
That would explain the problem if they are configured to have the same name.
What should I be looking for in man systemd.link?
And please do it onhttps://susepaste.org/, not on some commercial site. That is my own server actually. I can see it from here :)
I never heard of susepaste.org
I am going to confuse the matter more now. I just rebooted the system and now the two NICs are reversed. The primary NIC is now called rename2 and is no longer working. The second NIC which has not worked until no is now called eno1 and is working fine. :?
That just enforces theory that it is firmware problem. Provide information I asked for here and in another reply in this thread. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 19:17, Andrei Borzenkov wrote:
That just enforces theory that it is firmware problem. Provide information I asked for here and in another reply in this thread.
I assume you mean the following? lspci -nn grep -r . /sys/class/net/*/device/index Will send shortly. I have just done a few things in the meantime: I booted ubuntu 18.04 desktop live USB and I have eno1 and eno2 Both NICs appear to be working fine.... I have updated the BIOS to the latest one from Intel. I have reset the CMOS and configured BIOS settings again. I have reformatted the hard disk and I am currently re-installing ubuntu 18.04.1 from the latest ISO release. Let's see what happens after it is done. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Andrei, you were correct in thinking it might have been firmware related. After flashing the BIOS (which included the firmware for the network controllers). The second NIC is now recognised as eno2 and both are working perfectly. The BIOS I flashed was a new download of the same version already installed. I have done some more testing and discovered that the BIOS that was installed on the board was modified by the OEM (I assume in order to put their logo on it). I re-flashed this version and the problem came back. So I can only assume they messed up somewhere. However I have decided to update it to the latest BIOS as recommended on Intel's release notes and all is well. What a pain in the unmentionables! As soon as I re-read your message Andrei where you mentioned Firmware I though hmmm.. and it clicked! I remember a similar problem at my old job. Looking at my old notes I had almost the exact same problem. The board was an Intel DH61DL and the network controller did not get detected by the OS and was called eth0 not enp0s0 (OpenSUSE 12.3). The fix was BIOS update. The funny thing is that this board was from the exact same OEM (OEGStone) Stone Computers. They supplied the computers to my old employer. However this is quite a coincidence seeing as I bought this machine second hand over 3 years later. Makes you think right...? Anyway thank you all for your help. I have learned a lot on this one. Next job... :) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Another little quirk with the older intel firmware for this board that might make you chuckle. The setting uefi_enable can be set to 0 or 1 as you might expect But the BIOS firware is backward. When it says UEFI Enabled it sets uefi_enable to 0 and vice versa.. Confused me for a bit but they have fixed it in the later one at least. I might stick to ASUS boards in future. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-10-23 22:45, Paul Groves wrote:
But the BIOS firware is backward. When it says UEFI Enabled it sets uefi_enable to 0 and vice versa..
Kind of reminds me of my time as a VLSI engineer. We used a french software for construction of CPU's and in the software there was this dotted grid on the canvas you could turn on and off. The settings for turning it on was "No dots off". Still amuses me :) Cheers, -- /bengan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 22/10/2018 à 19:29, Paul Groves a écrit :
The system is running ubuntu 18.04. Now I have probably confused you by
may be try any Leap 15.0 live so to be more familiar for us and see if the problem is still there jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 22/10/2018 19:29, Paul Groves wrote:
I have never seen a NIC called rename before. What is this trickery?!?!?
I have not seen any post from you on the Ubuntu mailing list.
From a quick Google, it seems to be a known problem and has been for years. Various fixes are described.
https://askubuntu.com/questions/354783/why-is-network-interface-loaded-as-re... https://ubuntuforums.org/showthread.php?t=2207693 https://unix.stackexchange.com/questions/347735/network-interfaces-rename-no... Do any of those help? -- Liam Proven - Technical Writer, SUSE Linux s.r.o. Corso II, Křižíkova 148/34, 186-00 Praha 8 - Karlín, Czechia Email: lproven@suse.com - Office telephone: +420 284 241 084 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 10:34, Carlos E. R. wrote:
On 23/10/2018 11.30, Liam Proven wrote:
I have not seen any post from you on the Ubuntu mailing list.
He used the forum (got no response), and wondered here if there is a mail list there.
Exactly what I said yesterday. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
The following is from syslog after I run sudo netplan apply to re-apply configuration. I notice it is now renamed to rename5 also this line stands out: Oct 23 16:01:26 media systemd-udevd[475]: seq 4657 '/devices/pci0000:00/0000:00:1c.6/0000:03:00.0/net/eth0' is taking a long time syslog: Oct 23 16:00:26 media kernel: [ 474.175961] e1000e 0000:03:00.0 rename4: removed PHC Oct 23 16:00:26 media systemd-udevd[1742]: error changing net interface name 'eth0' to 'eno1': No such device Oct 23 16:00:26 media systemd-udevd[1742]: could not rename interface '4' from 'eth0' to 'eno1': No such device Oct 23 16:00:26 media kernel: [ 474.227652] e1000e 0000:03:00.0: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode Oct 23 16:00:26 media systemd-networkd[1764]: eno1: Gained IPv6LL Oct 23 16:00:26 media systemd-networkd[1764]: Enumeration completed Oct 23 16:00:26 media systemd-networkd[1764]: lo: Link is not managed by us Oct 23 16:00:26 media kernel: [ 474.281070] e1000e 0000:03:00.0 0000:03:00.0 (uninitialized): registered PHC clock Oct 23 16:00:26 media kernel: [ 474.334367] e1000e 0000:03:00.0 eth0: (PCI Express:2.5GT/s:Width x1) 7c:05:07:0f:11:c5 Oct 23 16:00:26 media kernel: [ 474.334370] e1000e 0000:03:00.0 eth0: Intel(R) PRO/1000 Network Connection Oct 23 16:00:26 media kernel: [ 474.334460] e1000e 0000:03:00.0 eth0: MAC: 3, PHY: 8, PBA No: FFFFFF-0FF Oct 23 16:00:26 media kernel: [ 474.336856] e1000e 0000:03:00.0 rename5: renamed from eth0 Oct 23 16:00:26 media dbus-daemon[967]: [system] Activating via systemd: service name='org.freedesktop.hostname1' unit='dbus-org.freedesktop.hostname1.service' requested by ':1.12' (uid=100 pid=1764 comm="/lib/systemd/systemd-networkd " label="unconfined") Oct 23 16:00:26 media systemd-networkd[1764]: eno1: DHCPv4 address 192.168.0.164/24 via 192.168.0.1 Oct 23 16:00:26 media systemd-networkd[1764]: eth0: Interface name change detected, eth0 has been renamed to rename5. Oct 23 16:00:26 media systemd-networkd[1764]: eno1: Configured Oct 23 16:01:26 media systemd-udevd[475]: seq 4657 '/devices/pci0000:00/0000:00:1c.6/0000:03:00.0/net/eth0' is taking a long time Oct 23 16:01:56 media systemd-udevd[1787]: error changing net interface name 'eth0' to 'eno1': File exists Oct 23 16:01:56 media systemd-udevd[1787]: could not rename interface '5' from 'eth0' to 'eno1': File exists Oct 23 16:01:56 media networkd-dispatcher[1038]: WARNING:Unknown index 5 seen, reloading interface list -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 10:30, Liam Proven wrote:
https://askubuntu.com/questions/354783/why-is-network-interface-loaded-as-re...
https://ubuntuforums.org/showthread.php?t=2207693
https://unix.stackexchange.com/questions/347735/network-interfaces-rename-no...
Do any of those help?
No really. I am not sure I understand all of it. It seems to me that those people are trying to rename interfaces though? I am not, this is just a default clean install -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Paul Groves wrote:
On 23/10/2018 10:30, Liam Proven wrote:
https://askubuntu.com/questions/354783/why-is-network-interface-loaded-as-re...
https://unix.stackexchange.com/questions/347735/network-interfaces-rename-no...
Do any of those help?
No really. I am not sure I understand all of it. It seems to me that those people are trying to rename interfaces though? I am not, this is just a default clean install
IIRC, the renaming happens as part of normal boot-up. At first, the devices are enumerated by the kernel/drivers - eth0,1,2,3,4,5,6,7,8,9. If you want 'predictable names', such as e.g. 'enp2s0', a renaming takes place. If you want to keep the ethX names, a renaming still takes place, as the initial kernel enumeration may not coincide with what you want. I'm sure we (openSUSE) also used names such as 'renameX' at some point, but I'm not aware of the current process. -- Per Jessen, Zürich (13.5°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 17:19, Per Jessen wrote:
IIRC, the renaming happens as part of normal boot-up. At first, the devices are enumerated by the kernel/drivers - eth0,1,2,3,4,5,6,7,8,9. If you want 'predictable names', such as e.g. 'enp2s0', a renaming takes place. If you want to keep the ethX names, a renaming still takes place, as the initial kernel enumeration may not coincide with what you want. I'm sure we (openSUSE) also used names such as 'renameX' at some point, but I'm not aware of the current process.
OK that makes much more sense when you put it like that. It looks like the Kernel is enumerating the adapter correctly because as shown in the log I attached previously it has eth0 and eth1 then it shows it renames eth0 to eno1 then it fails to rename eth1 so it gets stuck with it's temporary rename3 Had a look at another machine to see how it names the NICs. Ubuntu 18.04 names the interfaces on that machine as follows: eno1 enp2s0 enp2s1 enp3s0 I believe those represent that machines hardware which is: Onboard LAN Dual Lan PCIe card Single LAN PCI Card so on this PC which has two onboard NICs I expect it should show something like: eno1 enp2s0 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
23.10.2018 20:16, Paul Groves пишет:
On 23/10/2018 17:19, Per Jessen wrote:
IIRC, the renaming happens as part of normal boot-up. At first, the devices are enumerated by the kernel/drivers - eth0,1,2,3,4,5,6,7,8,9. If you want 'predictable names', such as e.g. 'enp2s0', a renaming takes place. If you want to keep the ethX names, a renaming still takes place, as the initial kernel enumeration may not coincide with what you want. I'm sure we (openSUSE) also used names such as 'renameX' at some point, but I'm not aware of the current process.
OK that makes much more sense when you put it like that.
It looks like the Kernel is enumerating the adapter correctly because as shown in the log I attached previously it has eth0 and eth1
then it shows it renames eth0 to eno1 then it fails to rename eth1 so it gets stuck with it's temporary rename3
Had a look at another machine to see how it names the NICs. Ubuntu 18.04 names the interfaces on that machine as follows: eno1 enp2s0 enp2s1 enp3s0
I believe those represent that machines hardware which is: Onboard LAN Dual Lan PCIe card Single LAN PCI Card
so on this PC which has two onboard NICs I expect it should show something like: eno1 enp2s0
Please show output of lspci -nn grep -r . /sys/class/net/*/device/index -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 17:44, Paul Groves wrote:
No really. I am not sure I understand all of it. It seems to me that those people are trying to rename interfaces though? I am not, this is just a default clean install
Um. Did you read them through to the end? They all contain resolutions, AFAICS. That's why I posted them. Exactly as Per K said -- the interface _normally_ gets renamed during bootup, to a temporary name and then a final one. Yours is getting stuck on the temporary one. That's what those links were saying and offering solutions to. My point about the Ubuntu mailing list is that I would have thought it was a more logical place to ask than the openSUSE list, no? -- Liam Proven - Technical Writer, SUSE Linux s.r.o. Corso II, Křižíkova 148/34, 186-00 Praha 8 - Karlín, Czechia Email: lproven@suse.com - Office telephone: +420 284 241 084 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 17:23, Liam Proven wrote:
My point about the Ubuntu mailing list is that I would have thought it was a more logical place to ask than the openSUSE list, no?
Yes definitely. Where can I subscribe to the Ubuntu mailing list? I filled out this several times but I never receive an email. Not even in my Spam folder. https://lists.ubuntu.com/mailman/listinfo/ubuntu-users Is this the right place? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 19:13, Paul Groves wrote:
I filled out this several times but I never receive an email. Not even in my Spam folder. https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
Is this the right place?
I've just checked and yes, that is the one that I used, way back when. It still recognises me. Might be worth trying to contact Ubuntu some other way and saying that it's broken...? -- Liam Proven - Technical Writer, SUSE Linux s.r.o. Corso II, Křižíkova 148/34, 186-00 Praha 8 - Karlín, Czechia Email: lproven@suse.com - Office telephone: +420 284 241 084 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Paul Groves wrote:
On 23/10/2018 17:23, Liam Proven wrote:
My point about the Ubuntu mailing list is that I would have thought it was a more logical place to ask than the openSUSE list, no?
Yes definitely.
More logical perhaps, but if the response is better here, why not? We are an inclusive community, not exclusive.
Where can I subscribe to the Ubuntu mailing list?
I filled out this several times but I never receive an email. Not even in my Spam folder. https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
Is this the right place?
Looks good to me. -- Per Jessen, Zürich (12.8°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Per Jessen wrote:
Paul Groves wrote:
I filled out this several times but I never receive an email. Not even in my Spam folder. https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
Is this the right place?
Looks good to me.
I have just subscribed, worked fine. The archive indicates the list isn't perhaps overly active. -- Per Jessen, Zürich (12.8°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 23/10/2018 19:45, Per Jessen wrote:
I have just subscribed, worked fine. The archive indicates the list isn't perhaps overly active.
It's not any more. It comes and goes. Sadly, Ubuntu killed their chat list (ubuntu-sounder) years ago, and this has caused the mailing-list community to die off somewhat. -- Liam Proven - Technical Writer, SUSE Linux s.r.o. Corso II, Křižíkova 148/34, 186-00 Praha 8 - Karlín, Czechia Email: lproven@suse.com - Office telephone: +420 284 241 084 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (7)
-
Andrei Borzenkov -
Bengt Gördén -
Carlos E. R. -
jdd＠dodin.org -
Liam Proven -
Paul Groves -
Per Jessen