vmfex - veth "unknown e 1"

NaelShahid_2 · ‎09-05-2012

I have an issue in which a VM vnic does not show up within the VMs tab within UCSM, the veth on the FI also lists the interface status as ‘unknown e 1’. Note this is not high performance vm-fex.

What I have found is the issue only arises when the VM is attached to dv port 2013 and corresponding veth32769.

I have moved VMs between port profiles, hosts and failed between fabrics to rule out config / hardware, in the end it always boils down to dv port 2013 and veht32769.

This is a new install running 2.0(3b) with vsphere 5.0.

See interface detail below:

xxxxxx(nxos)# show interface Veth32769

Vethernet32769 is down (vicEnNotRcvd)

Bound Interface is port-channel1283

Hardware: Virtual, address: 547f.ee98.e0e0 (bia 547f.ee98.e0e0)

Encapsulation ARPA

Port mode is access

EtherType is 0x8100

Rx

0 unicast packets 0 multicast packets 0 broadcast packets

0 input packets 0 bytes

0 input packet drops

Tx

0 unicast packets 0 multicast packets 0 broadcast packets

0 output packets 0 bytes

0 flood packets

0 output packet drops

xxxxxx(nxos)# show interface Veth32769 detail

vif_index: 270

--------------------------

veth is bound to interface port-channel1283 (0x16000502)

priority: 0

vntag: 0

status: active

registered mac info:

xxxxxxxx(nxos)# show interface port-channel 1283

port-channel1283 is up

Hardware: Port-Channel, address: a493.4c64.bb64 (bia a493.4c64.bb64)

MTU 1500 bytes, BW 20000000 Kbit, DLY 10 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation ARPA

Port mode is vntag

full-duplex, 10 Gb/s

Beacon is turned off

Input flow-control is off, output flow-control is on

Switchport monitor is off

EtherType is 0x8100

Members in this channel: Eth1/1/1, Eth1/1/3

Last clearing of "show interface" counters never

30 seconds input rate 187808 bits/sec, 23476 bytes/sec, 12 packets/sec

30 seconds output rate 125984 bits/sec, 15748 bytes/sec, 7 packets/sec

Load-Interval #2: 5 minute (300 seconds)

input rate 51.70 Kbps, 2 pps; output rate 39.41 Kbps, 3 pps

RX

30381043 unicast packets 55534 multicast packets 2311 broadcast packets

30438888 input packets 57289907672 bytes

26696617 jumbo packets 0 storm suppression bytes

0 giants 0 input error 0 short frame 0 overrun 0 underrun 0 watchdog 0 if down drop

0 input with dribble 0 input discard

0 Rx pause

TX

20560423 unicast packets 216192 multicast packets 17387 broadcast packets

20794002 output packets 32060343014 bytes

14783405 jumbo packets

0 output errors 0 collision 0 deferred 0 late collision

0 lost carrier 0 no carrier 0 babble

0 Tx pause

18 interface resets

abbharga · ‎09-05-2012

Hi Nael,

Can you help me with a screen shot from the UCSM where you see the interface as unknown?

Looking at the interface counters i would assume the VM is able to communicate fine..correct?

What type of adapter is it e1000 or vmxnet3?

./Abhinav

NaelShahid_2 · ‎09-06-2012

Within UCSM the vnic does not show at all so there is nothing I can show from here. The output of “show interface status” is listed within initial post.

Today I removed the dvs / port profiles and recreated. The very first vm I attached to a port group did not show within UCSM, looking within nxos again the first veth is 32769 and again it lists a status of “unknown e 1”…. So this is not specific to the dv port, but only the veth port of 32769.

Note this can happen to any vm which gets mapped to veth 32769 and they all use vmxnet3 adapters………….

abbharga · ‎09-06-2012

Hi Nael,

I am trying to put a logic here to see if this is a new bug or something, are there any VMs which use veth number above 32768?

./Abhinav

NaelShahid_2 · ‎09-06-2012

Hi

Yes, there certainly are… In fact it looks as tho 32768 is the very first veth assigned to VMs, all other ports start directly after this number… See below:

Veth9819 server 1/1, VHBA v connected 950 auto auto --

Veth9931 server 1/5, VHBA v nonPcpt 950 auto auto --

Veth32769 -- unknown e 1 auto auto --

Veth32770 -- connected trunk auto auto --

Veth32771 -- connected trunk auto auto --

abbharga · ‎09-06-2012

Hi Nael,

Can you do one more thing:

1) from the local-mgmt of the primary FI, run the following log capture:

#tail-mgmt-log svc_sam_extvmmAG

and try to assing this port to a VM, this should give us some idea on whats happening.

./Abhinav

NaelShahid_2 · ‎09-06-2012

Yes no problem. Once I attach the VM how to I get the log file? (or is it console output?)

abbharga · ‎09-06-2012

It is a console output, however you maywant to log the ssh session to a file.

./Abhinav

NaelShahid_2 · ‎09-06-2012

I did this, however no output was displayed..?

abbharga · ‎09-06-2012

can you check on the other FI?

I am looking for any known issue around this, if i don't find anything you can go ahead and open a TAC case and we can take a look at the logs etc from there.

./Abhinav

NaelShahid_2 · ‎09-06-2012

Check both FIs, still not outputting, even with other ports which are fine.

PS: This system does not have support - Only warranty..

NaelShahid_2 · ‎09-06-2012

do you want me to try something else?

regards