VMware Cloud Community
ioahdsf8048ygh
Contributor
Contributor

esxi 6.5 vmnic with "no networks"

3x Fujitsu Primergy RX2510 M2

running ESXi 6.5 u1 (Fujitsu VMvisor-Installer-6.5-5969303-v411-1) with Intel I350 4port NICs

All drivers are as vmware/Fujitsu ISO standard.

HP ProCurve Switch 5406zl  with cluster uplinks connected to the same 24 port Gig-T ZL module.

vswitch.pngvmnics.png

all 4 uplinks on each port group are patched into the same switch, have the same switch port configs (untagged vlan xxx)

I have an intermittent issue with 4 vmnic4 showing "no networks". The vmnic in question is one port on a 4 port Intel I350 P4 card.

Once a vmnic shows this state, its stuck that way until I either down and up the vmnic, or reboot the host.

A week later a second host did exactly the same with a different port on the same make/model of 4-port card.

about 10 days later the 3rd host also did the same with a different port again.

If i vmotion the VM to another host its fine.

If i drop the problem vmnic out of the port group, the vms are fine on the remaining uplinks.

the problem NIC on each server is one port in a 4 port card, with all 3 other links functioning perfectly before and after.

I have been all over this including wiresharking the problem port from a VM.  The "no networks" uplink still shows some traffic from the LAN, however a lot of ARPs are going unanswered etc. so Im not convinced this is a flakey port.  However Ive also replaced the 4 port NIC in one server, only to have the issue reappear a week later.

esxcli network nic stats get -n VMNIC4 reveal the following:

Receive FIFO errors 57039176

Receive missed errors: 57039176

No other errors are shown.

Any pointers on this would be a huge help!

Cheers

O

Reply
0 Kudos
12 Replies
msripada
Virtuoso
Virtuoso

Check the CDP for the vmnic and compare with other vmnic -> validate the configuration on physical switch with other vmnic.

I think but am not sure that this may appear if the NIC never observed or used any traffic.

Thanks,

MS

Reply
0 Kudos
daphnissov
Immortal
Immortal

What version is that igbn driver?

Reply
0 Kudos
IT_pilot
Expert
Expert

how many network cards do you have on one host?

http://it-pilot.ru
Reply
0 Kudos
ioahdsf8048ygh
Contributor
Contributor

Hi,

Show CDP Nei on the switch returns nothing for that port.

The uplink has been active previously however not for days so suspect anything which would have been listed here has long since aged out.

Other ports on the switch do list device IDs as expected.

running esxcli network ip neighbor list on the host side lists neighbors correctly for all other vmknics but nothing under that uplink.

Reply
0 Kudos
ioahdsf8048ygh
Contributor
Contributor

Hi,

I have 2x cards per host, all hosts are the same.

1x 2 port card, and 1x 4 port card.

Both are Intel I350, the 2 port being LOM, the 4 port being added.

Cheers

Reply
0 Kudos
ioahdsf8048ygh
Contributor
Contributor

Hi,

This is as standard with the Fujitsu OEM build.

Driver IGBN

Firmware 1.63.0:0x80000ee3

Version 1.3.1

Cheers

Reply
0 Kudos
ioahdsf8048ygh
Contributor
Contributor

OK, so in a completely unrelated site / cluster, I have now seen the same issue on ESXi 5.5 host with the same I350 4x port cards.

Has anyone else seen this issue?

Simply downing and upping the port with esxcli resolved the issue briefly, for a day to a week, then the issue returns...

I also still have this issue remaining on the original cluster.

I have tickets open with both VMware and Fujitsu however neither are particularly interested or progressing well...

Reply
0 Kudos
Virtguy
Enthusiast
Enthusiast

Same issue .. EXACT issue here

with 6.5.0 5969303

fujitsu rx2510 M2

intel i350

any news ?

Reply
0 Kudos
rollerone11
Contributor
Contributor

Same issus and trouble...
with 6.5.0.5.9.73321
Same HP procurve

Any news about a solution ?

Reply
0 Kudos
kenbshinn
Enthusiast
Enthusiast

I have had issues like this before. It is usually a result of not having the port on the physical network switch enabled with the correct vlans.

I did see you mention

If i vmotion the VM to another host its fine.

What VM are you referring to? The VCSA Appiliance? Any VM?

Reply
0 Kudos
CJPM
Contributor
Contributor

Same issus
with VMware ESXi, 6.7.0, 10764712

FUJITSU PRIMERGY RX2530 M2
and Intel Corporation I350 Gigabit Network Connection

Does anyone have a solution?

Reply
0 Kudos
Virtguy
Enthusiast
Enthusiast

So, seems to me the problem is :

Fujitsu rx + intel i350 + >esxi 6.5.0

In other situations i discovered bugs and problems in firmwares or hardware, and its very dificult to translate to fabricant....

But if every single one of us.. open a ticket.. and redirect to this post... it will help...

More information..

the problem appears one time ( only ) after restart... if i change vmnic affected and put out of vswitch. never appears again.. until i restart or the servers power off and on.

Seems to me the network card in system is unable to maintain 4 ports workings.. and randomly one of the four ports ( one diferent every time ) is affected by this use.

Reply
0 Kudos