Public bug reported:
With linux-hwe-generic-trusty we see kern.log entries for "TX driver
issue detected, PF reset issued" on a NIC in a bond, followed by the
same interface going down and up again. After this time networking to
virtual machines bridged to that bond fails.
We see either asymmetric traffic, with ARP replies not reaching the VMs,
or no traffic at all to or from VMs.
In this context, these are nova-compute nodes using OpenvSwitch and KVM.
Reloading the i40e module corrects the networking.
Setting TCP Segmentation Offload off on the NICs seems to prevent the PF
reset, and we don't believe we have seen VM networking problems since
doing so.
# ethtool -i eth5
driver: i40e
version: 1.3.4-k
firmware-version: f4.33.31377 a1.2 n4.41 e1866
bus-info: 0000:04:00.1
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes
supports-priv-flags: yes
** Affects: linux (Ubuntu)
Importance: Undecided
Status: New
** Tags: canonical-bootstack
** Package changed: linux-lts-vivid (Ubuntu) => linux-meta (Ubuntu)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1540405
Title:
i40e fails virtual networking
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1540405/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
