[SR-IOV]: Use static MAC addresses with udev rules for renaming VF interfaces (Mellanox CX4) (Documentation)

Bug #1689616 reported by bugproxy
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
The Ubuntu-power-systems project
Invalid
Low
Breno Leitão
ubuntu-docs (Ubuntu)
Invalid
Low
Ubuntu on IBM Power Systems Bug Triage

Bug Description

Problem Description
=============
The udev rules for renaming the SR-IOV Mellanox CX4 VF interface are not getting
implemented. After setting the udev rules, and unbind/bind the drivers for the VF's,
the interface name still doesn't change. Same issue isn't noticed for the PF of the
SR-IOV adapter.

Resolution
=======
Use static MAC addresses for renaming SRIOV Mellanox CX4 VF interfaces with udev rules.

Documentation
==========
With Mellanox CX4 SR-IOV when the VFs are created, each of them will get a random MAC address.
To be able to use the '70-persistent-net.rules' to define interfaces based on MAC Address, one need
to set static MAC address to the VF.

This is an example
SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="02:a9:bb:cc:dd:00", NAME="enP4p1s0vf"

First execute below command
echo 02:a9:bb:cc:dd:00 > /sys/class/infiniband/mlx5_0/device/sriov/0/mac

and then
do the unbind/bind
echo 0004:01:00.2 > /sys/bus/pci/drivers/mlx5_core/unbind
echo 0004:01:00.2 > /sys/bus/pci/drivers/mlx5_core/bind

then one can see
enP4p1s0vf: flags=4098<BROADCAST,MULTICAST> mtu 1500
        ether 02:a9:bb:cc:dd:00 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

bugproxy (bugproxy)
tags: added: architecture-ppc64le bugnameltc-153427 severity-low targetmilestone-inin1710
Changed in ubuntu:
assignee: nobody → Ubuntu on IBM Power Systems Bug Triage (ubuntu-power-triage)
affects: ubuntu → ubuntu-docs (Ubuntu)
Revision history for this message
bugproxy (bugproxy) wrote : Comment bridged from LTC Bugzilla

------- Comment From <email address hidden> 2017-05-09 14:14 EDT-------
Hello Canonical,

Please include above documentation text for using static MAC addresses
for renaming SR-IOV Mellanox CX4 VF interfaces with udev rules.

Revision history for this message
Gunnar Hjalmarsson (gunnarhj) wrote :

ubuntu-docs is a package for the desktop documentation, and this issue is apparently not related to that. Which set of docs is it you would like to see updated?

Changed in ubuntu-docs (Ubuntu):
status: New → Incomplete
Manoj Iyer (manjo)
tags: added: ubuntu-17.04
Manoj Iyer (manjo)
Changed in ubuntu-docs (Ubuntu):
importance: Undecided → Low
Changed in ubuntu-power-systems:
importance: Undecided → Low
Manoj Iyer (manjo)
Changed in ubuntu-release-notes:
assignee: nobody → Ubuntu on IBM Power Systems Bug Triage (ubuntu-power-triage)
Manoj Iyer (manjo)
Changed in ubuntu-docs (Ubuntu):
status: Incomplete → Invalid
Changed in ubuntu-release-notes:
assignee: Ubuntu on IBM Power Systems Bug Triage (ubuntu-power-triage) → Canonical Foundations Team (canonical-foundations)
tags: added: triage-g
Manoj Iyer (manjo)
tags: added: triage-a
removed: triage-g
Revision history for this message
Steve Langasek (vorlon) wrote :

I don't find this request very clear. Several questions:
 - What release are you seeking to have this release noted for? tags suggest 17.04 or 17.10 or both, but neither of these are LTS releases, so I have to question whether the tags are correct.
 - Is the lack of persistent mac addresses a problem with the kernel driver? With the underlying hardware?
 - In what environment should we expect to see this problem? Is it POWER-specific?
 - The workaround suggested appears to be a one-off change, that would still not persist across reboots of the system (or perhaps the state persists within the VM, but would not persist across a cold boot). Have I understood correctly? If so, this does not seem like a complete workaround that warrants recommending to users in the release notes.
 - Why is this being suggested as a release notes change, instead of a request for a fix in the implementation?

Changed in ubuntu-release-notes:
status: New → Incomplete
Frank Heimes (fheimes)
Changed in ubuntu-power-systems:
status: New → Incomplete
Revision history for this message
Manoj Iyer (manjo) wrote :

Breno,

Could you please help answer Steve's questions ?

Changed in ubuntu-power-systems:
assignee: nobody → Breno Leitão (breno-leitao)
Manoj Iyer (manjo)
tags: added: triage-g
removed: triage-a
Revision history for this message
bugproxy (bugproxy) wrote :

------- Comment From <email address hidden> 2017-12-04 11:28 EDT-------
Since we are not seeing update on this bug, and neither I or Carol knows where this bug should go, I am rejecting/closing it.

Revision history for this message
Andrew Cloke (andrew-cloke) wrote :

Thanks Breno. I'll close out the LP bug.

Changed in ubuntu-power-systems:
status: Incomplete → Invalid
no longer affects: ubuntu-release-notes
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.