User Activity

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    I have insured ALL devices are at 1500 MTU. The rx_long_byte_count counter on both ports is still incrementing after a reboot. Are you sure it stands for invalid MTUs? I have even put the switch at only allowing 1500 MTU. After disabling traffic shapping and my IPv6-to-4 tunnel the issue occurred again today... :( I have now disabled C-STATES in the BIOS to see if it is a CPU/hardware issue. I will try the kernel list next.

  • Modified a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    I changed the p4p2 adapter MTU to 1500. However, the issue has occurred again. I am attaching the full dmesg and ethtool output for -k and -S. I saw rx_long_byte_count is still present, but I believe this is due to my tun6to4 adapter (IPv6 in IPv4) which has an MTU of 1490 (10-bytes of tunnel header overhead). My next steps will be to disable the tun6to4 adapter as well as my traffic shapper setup with tc queues. Any other advice is appreciated. Should I ping the kernel mailing list with this bu...

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    I changed the p4p2 adapter MTU to 1500. However, the issue has occurred again. I am attaching the full dmesg, and ethtool output for -k and -S. I saw rx_long_byte_count is still present, but I believe this is due to my tun6to4 adapter (IPv6 in IPv4) which has an MTU of 1490 (10-bytes of tunnel header overhead). My next steps will be to disable the tun6to4 adapter as well as my traffic shapper setup with tc queues. Any other advice is appreciated. Should I ping the kernel mailing list with this b...

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    Sorry, I will include the full kernel output for the last time the issue occured. The first ethtool -S output for p4p2 will show xon/xoff. The kernel also reported it: kernel: igb 0000:02:00.1 p4p2: igb: p4p2 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX and after I disabled it on the switch the kernel reported the change: kernel: igb 0000:02:00.1 p4p2: igb: p4p2 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX The MTU on p4p1 is 1500 and the MTU on p4p2 is 9000. The server is acting...

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    Disabling flow control did not help. The issue returned in less than 24 hours. I am attaching a zip file of the dmesg output and ethtool output.

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    The switch did have "Global Flow Control (IEEE 802.3x) Mode" Enabled. I have now disabled it and am running the I350 system on kernel 4.12. FYI: The I218-V and 82579V devices are running kernel 4.12 and have over a week of uptime with no network issues.

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    Here is ethtool -S for p4p2.

  • Posted a comment on ticket #574 on Intel Ethernet Drivers and Utilities

    The issue occurred twice this weekend. After 36 hours on kernel 4.11. After 1 hour on kernel 4.12. I am attaching ethtool -S output for p4p1. The next post will be for p4p2.

View All

Personal Data

Username:
vegetto34
Joined:
2002-04-29 04:05:30

Projects

  • No projects to display.

Skills

  • No skills entered.

Personal Tools