[CentOS] CentOS7 sometimes don't detect NIC after reboot

On Tue, Sep 17, 2019 at 9:11 AM Miroslav Geisselreiter <mg at intar.cz> wrote:
>
> Dne 17.9.2019 v 14:02 Mauricio Tavares napsal(a):
> > On Tue, Sep 17, 2019 at 7:06 AM Miroslav Geisselreiter <mg at intar.cz> wrote:
> >> I have brand new PC with this components:
> >> CPU Intel® Pentium G5400, LGA1151
> >> motherboard ASUS PRIME B360M-C
> >> 16 GB RAM
> >> HDD 2x ADATA SSD 256GB XPG GAMMIX S11, PCIe Gen3x4 M.2 2280 (RAID1)
> >> NIC Intel X550-T1 Ethernet Converged Network Adapter
> >>
> >> I installed CentOS 7 and two NICs were detected:
> >> eno1 (on motherboard)
> >> enp1s0 (Intel X550-T1)
> >>
> >> When I restart the machine sometimes enp1s0 is missing. It is not
> >> detected during boot. It looks like NIC card is not installed / not
> >> present. After next reboot everything is fine and I do not see any
> >> problems with NIC card.
> >>
> >> How can I avoid this problem with missing NIC? Can you help me, please?
> >>
> >        Nothing exciting on dmesg? Did you check the pci chain to see if
> > it is being reported as there?
> >
>  From /var/log/dmesg, when NIC is detected and working:
> grep ixgbe dmesg
> [    0.999078] ixgbe: Intel(R) 10 Gigabit PCI Express Network Driver -
> version 5.1.0-k-rh7.6
> [    0.999081] ixgbe: Copyright (c) 1999-2016 Intel Corporation.
> [    1.720556] ixgbe 0000:01:00.0: irq 138 for MSI/MSI-X
> [    1.720563] ixgbe 0000:01:00.0: irq 139 for MSI/MSI-X
> [    1.720568] ixgbe 0000:01:00.0: irq 140 for MSI/MSI-X
> [    1.720573] ixgbe 0000:01:00.0: irq 141 for MSI/MSI-X
> [    1.720578] ixgbe 0000:01:00.0: irq 142 for MSI/MSI-X
> [    1.720602] ixgbe 0000:01:00.0: Multiqueue Enabled: Rx Queue count =
> 4, Tx Queue count = 4 XDP Queue count = 0
> [    1.821743] ixgbe 0000:01:00.0: PCI Express bandwidth of 32GT/s available
> [    1.821754] ixgbe 0000:01:00.0: (Speed:8.0GT/s, Width: x4, Encoding
> Loss:N/a)
> [    1.933432] ixgbe 0000:01:00.0: MAC: 4, PHY: 0, PBA No: H92506-004
> [    1.933434] ixgbe 0000:01:00.0: a0:36:9f:f8:46:28
> [    2.093020] ixgbe 0000:01:00.0: Intel(R) 10 Gigabit Network Connection
>
> When NIC is not detected, there is no ixgbe message in dmesg.
>
> I can list irq from dmesg (with NIC detected) and dmesg.old (NIS is not
> tedected)
> grep irq dmesg
> [    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
> [    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 high level)
> [    0.000000] NR_IRQS:327936 nr_irqs:1024 0
> [    0.531915] pcieport 0000:00:01.0: irq 120 for MSI/MSI-X
> [    0.532038] pcieport 0000:00:1b.0: irq 121 for MSI/MSI-X
> [    0.532309] pcieport 0000:00:1c.5: irq 122 for MSI/MSI-X
> [    0.532534] pcieport 0000:00:1d.0: irq 123 for MSI/MSI-X
> [    0.604204] 00:05: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
> [    0.625761] 00:06: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
> [    0.628502] xhci_hcd 0000:00:14.0: irq 124 for MSI/MSI-X
> [    0.630583] i8042: PNP: PS/2 Controller [PNP0303:PS2K,PNP0f03:PS2M]
> at 0x60,0x64 irq 1,12
> [    0.634280] serio: i8042 KBD port at 0x60,0x64 irq 1
> [    0.634284] serio: i8042 AUX port at 0x60,0x64 irq 12
> [    0.635917] rtc_cmos rtc_cmos: alarms up to one month, y3k, 114 bytes
> nvram, hpet irqs
> [    1.027014] e1000e 0000:00:1f.6: irq 125 for MSI/MSI-X
> [    1.028089] nvme 0000:02:00.0: irq 126 for MSI/MSI-X
> [    1.029367] nvme 0000:06:00.0: irq 127 for MSI/MSI-X
> [    1.130027] nvme 0000:02:00.0: irq 126 for MSI/MSI-X
> [    1.130033] nvme 0000:02:00.0: irq 128 for MSI/MSI-X
> [    1.130038] nvme 0000:02:00.0: irq 129 for MSI/MSI-X
> [    1.130041] nvme 0000:02:00.0: irq 130 for MSI/MSI-X
> [    1.130045] nvme 0000:02:00.0: irq 131 for MSI/MSI-X
> [    1.233841] nvme 0000:06:00.0: irq 127 for MSI/MSI-X
> [    1.233846] nvme 0000:06:00.0: irq 132 for MSI/MSI-X
> [    1.233851] nvme 0000:06:00.0: irq 133 for MSI/MSI-X
> [    1.233855] nvme 0000:06:00.0: irq 134 for MSI/MSI-X
> [    1.233858] nvme 0000:06:00.0: irq 135 for MSI/MSI-X
> [    1.378373] ahci 0000:00:17.0: irq 136 for MSI/MSI-X
> [    1.433203] ata1: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439100 irq 136
> [    1.433214] ata2: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439180 irq 136
> [    1.433226] ata3: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439200 irq 136
> [    1.433238] ata4: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439280 irq 136
> [    1.433250] ata5: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439300 irq 136
> [    1.433262] ata6: SATA max UDMA/133 abar m2048 at 0xa2439000 port
> 0xa2439380 irq 136
> [    1.438562] i915 0000:00:02.0: irq 137 for MSI/MSI-X
> [    1.720556] ixgbe 0000:01:00.0: irq 138 for MSI/MSI-X
> [    1.720563] ixgbe 0000:01:00.0: irq 139 for MSI/MSI-X
> [    1.720568] ixgbe 0000:01:00.0: irq 140 for MSI/MSI-X
> [    1.720573] ixgbe 0000:01:00.0: irq 141 for MSI/MSI-X
> [    1.720578] ixgbe 0000:01:00.0: irq 142 for MSI/MSI-X
> [    2.762813] parport0: PC-style at 0x378, irq 5 [PCSPP]
> [    2.931480] snd_hda_intel 0000:00:1f.3: irq 143 for MSI/MSI-X
>
> grep irq dmesg.old
> [    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
> [    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 high level)
> [    0.000000] NR_IRQS:327936 nr_irqs:1024 0
> [    0.531838] pcieport 0000:00:01.0: irq 120 for MSI/MSI-X
> [    0.531961] pcieport 0000:00:1b.0: irq 121 for MSI/MSI-X
> [    0.532189] pcieport 0000:00:1c.5: irq 122 for MSI/MSI-X
> [    0.532360] pcieport 0000:00:1d.0: irq 123 for MSI/MSI-X
> [    0.604543] 00:05: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
> [    0.626093] 00:06: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
> [    0.628860] xhci_hcd 0000:00:14.0: irq 124 for MSI/MSI-X
> [    0.630950] i8042: PNP: PS/2 Controller [PNP0303:PS2K,PNP0f03:PS2M]
> at 0x60,0x64 irq 1,12
> [    0.634787] serio: i8042 KBD port at 0x60,0x64 irq 1
> [    0.634792] serio: i8042 AUX port at 0x60,0x64 irq 12
> [    0.636314] rtc_cmos rtc_cmos: alarms up to one month, y3k, 114 bytes
> nvram, hpet irqs
> [    0.990007] nvme 0000:02:00.0: irq 125 for MSI/MSI-X
> [    0.991375] ahci 0000:00:17.0: irq 126 for MSI/MSI-X
> [    0.992906] nvme 0000:06:00.0: irq 127 for MSI/MSI-X
> [    1.092088] nvme 0000:02:00.0: irq 125 for MSI/MSI-X
> [    1.092095] nvme 0000:02:00.0: irq 128 for MSI/MSI-X
> [    1.092099] nvme 0000:02:00.0: irq 129 for MSI/MSI-X
> [    1.092103] nvme 0000:02:00.0: irq 130 for MSI/MSI-X
> [    1.092107] nvme 0000:02:00.0: irq 131 for MSI/MSI-X
> [    1.095722] ata1: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339100 irq 126
> [    1.095776] ata2: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339180 irq 126
> [    1.095782] ata3: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339200 irq 126
> [    1.095787] ata4: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339280 irq 126
> [    1.095798] ata5: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339300 irq 126
> [    1.095810] ata6: SATA max UDMA/133 abar m2048 at 0xa1339000 port
> 0xa1339380 irq 126
> [    1.096503] e1000e 0000:00:1f.6: irq 132 for MSI/MSI-X
> [    1.196872] nvme 0000:06:00.0: irq 127 for MSI/MSI-X
> [    1.196877] nvme 0000:06:00.0: irq 133 for MSI/MSI-X
> [    1.196882] nvme 0000:06:00.0: irq 134 for MSI/MSI-X
> [    1.196886] nvme 0000:06:00.0: irq 135 for MSI/MSI-X
> [    1.196889] nvme 0000:06:00.0: irq 136 for MSI/MSI-X
> [    1.436844] i915 0000:00:02.0: irq 137 for MSI/MSI-X
> [    2.351174] parport0: PC-style at 0x378, irq 5 [PCSPP]
> [    2.525337] snd_hda_intel 0000:00:1f.3: irq 138 for MSI/MSI-X
>
> I do not know what you mean "check the pci chain". My knowledge of
> kernel level is weak, sorry. What else can I do?
>
      The "checking the pci chain" argument is that sometimes the card
is there but is cheerfully ignored. From

> [    1.720556] ixgbe 0000:01:00.0: irq 138 for MSI/MSI-X

You know that it is in bus 1 slot 0, so try something like

lspci -s 01:00.0 -v

to see if it reports something there. Just in the odd case the pci
side of your computer knows there is a card there but the kernel can't
be bothered (flaky driver?). I am probably not using the right
terminology, and clearly expect someone to set me right, I've had
network cards that would show in the pci chain but not in dmesg. Other
thing you want to think about is that I've had cards that only worked
if were inserted in a specific slot for no reason whatsoever.

> Miroslav
>
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> https://lists.centos.org/mailman/listinfo/centos