vmxnet3 interface renamed on upgrade 9.3 > 9.10

[clarknova]

I just upgraded a 9.3.1 virtual machine to 9.10. The first problem apparent after a reboot is that one of my interfaces, a vmxnet3 adapter was renamed, causing the device to remain unconfigured on boot. Fortunately the system has an em0 NIC so I was able to log into the web UI and reconfigure the "new" NIC.

Not a huge problem here, but a definite caveat to anybody upgrading remotely.

 

[Mirfster]

Think that is due to 9.10 having those drivers in the kernel. Now if you would have just used the "E1000" it should have went fine. ;)

Also, there is always (should be at least) IPMI for those that are working on a machine remotely.

Side note: Those are not the most recent drivers and have been know to cause issues with jails (if you are using ESXi 6.0). See my posting near the bottom of this thread for info: https://forums.freenas.org/index.php?threads/my-dream-system-i-think.41072/page-11

 

[clarknova]

E1000 is gigabit only, so not a proper substitute.

 

[Ericloewe]

E1000 is gigabit only, so not a proper substitute.

That's apparently a misconception. I heard so from a green guy with a penchant for stealing toys.

 

[clarknova]

You might need to find yourself a new leprechaun then.

# dmesg | grep eth
[ 0.839110] vmxnet3 0000:03:00.0 eth0: NIC Link is Up 10000 Mbps
[ 0.876759] vmxnet3 0000:03:00.0 ens160: renamed from eth0
[ 1.188636] e1000 0000:02:01.0 eth0: (PCI:66MHz:32-bit) 00:0c:29:17:fd:b3
[ 1.188748] e1000 0000:02:01.0 eth0: Intel(R) PRO/1000 Network Connection
[ 1.190079] e1000 0000:02:01.0 ens33: renamed from eth0
# ethtool ens33
Settings for ens33:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supported pause frame use: No
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Advertised pause frame use: No
Advertised auto-negotiation: Yes
Speed: 1000Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 0
Transceiver: internal
Auto-negotiation: on
MDI-X: Unknown (auto)
Supports Wake-on: d
Wake-on: d
Current message level: 0x00000007 (7)
drv probe link
Link detected: no

 

[joeschmuck]

Lets put this into perspective...

The E1000 will run at higher internal traffic throughput within ESXi than 1Gb (I have seen it with my own eyes). The vmxnet3 driver will run much faster internal to ESXi than the E1000 nic driver. Neither one will exceed 1Gb/sec external to ESXi if the physical interface is 1Gb.

 

[clarknova]

Lets put this into perspective...

The E1000 will run at higher internal traffic throughput within ESXi than 1Gb (I have seen it with my own eyes). The vmxnet3 driver will run much faster internal to ESXi than the E1000 nic driver. Neither one will exceed 1Gb/sec external to ESXi if the physical interface is 1Gb.

So are you saying

1Gb < E1000 < vmxnet3

on the ESXi virtual switch?

 

[Spearfoot]

So are you saying

1Gb < E1000 < vmxnet3

on the ESXi virtual switch?

Yes - I regularly get nearly 3Gbps using the E1000 driver w/ a VMware virtual switch:

 

[joeschmuck]

So are you saying

1Gb < E1000 < vmxnet3

on the ESXi virtual switch?

Yes, but over a physical connection you will only get what the physical connection can provide.

 

[Mirfster]

I did some testing over here: "Interesting Tests with 10GB on ESXi 6.0 U2" and still need to do a little more. I still have more to do, but think that my issue may be the CPUs. Not sure yet though...