Portchannel question

[grapevine]

We have 3 Cisco IE-9320 in a stack. We have a Portchannel connecting to ESXI from switch 2 port 3 and switch 3 port 4

4 Po4(SU) - Gi2/0/3(P) Gi3/0/4(P).

Ports are in bundled.

GigabitEthernet2/0/3 media type is 10/100/1000BaseTX SFP

GigabitEthernet3/0/4  media type is 10/100/1000BaseTX 

ESXI nic teaming is configured as Route based on IP hash.

Ping to esxi is dropping intermittently

SW01#sh run int Po4
Building configuration...

Current configuration : 127 bytes
!
interface Port-channel4
switchport trunk allowed vlan 16,18
switchport mode trunk
end

SW01#sh run int Gi2/0/3
Building configuration...

Current configuration : 168 bytes
!
interface GigabitEthernet2/0/3
switchport trunk allowed vlan 16,18
switchport mode trunk
channel-group 4 mode on
end

SW01#sh run int Gi3/0/4
Building configuration...

Current configuration : 168 bytes
!
interface GigabitEthernet3/0/4
switchport trunk allowed vlan 16,18
switchport mode trunk
channel-group 4 mode on
end

arp is incomplete for esxi in firewalls. please advise how to fix it

[MHM Cisco World]

monitor capture PO type traffic interface Port-channel1 both

monitor capture PO match ipv4 host <EXSI host IP> any

monitor capture PO start <<- start ping to ESXI one for host in vlan16 and other in vlan 18

monitor capture PO stop

show monitor capture PO brief 

share the output of brief 

MHM

[grapevine]

PVW-CSW01#show monitor capture po parameter
monitor capture po interface Port-channel4 BOTH
monitor capture po match ipv4 any host 10.150.16.43
monitor capture po buffer size 10
monitor capture po limit pps 1000

PVW-CSW01#show monitor capture po ?
buffer Display captured buffer
parameter Display commands used to create capture point
| Output modifiers
<cr> <cr>

PVW-CSW01#show monitor capture po buffer
buffer size (KB) : 10240
buffer used (KB) : 0
packets in buf : 0
packets dropped : 0
packets per sec : 0

[MHM Cisco World]

did you  start ping then stop the capture ? 

PVW-CSW01#show monitor capture po buffer brief <<-

[grapevine]

Yes I started and stopped

PVW-CSW01#show monitor capture po parameter
monitor capture po interface Port-channel4 BOTH
monitor capture po match ipv4 any host 10.150.16.43
monitor capture po buffer size 10
monitor capture po limit pps 1000
PVW-CSW01#monitor capture po start
Started capture point : po
PVW-CSW01#ping 10.150.16.43
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.150.16.43, timeout is 2 seconds:
.....
Success rate is 0 percent (0/5)
PVW-CSW01#monitor capture po stop
Bytes dropped in asic - 0

Stopped capture point : po
PVW-CSW01#show monitor capture po parameter
monitor capture po interface Port-channel4 BOTH
monitor capture po match ipv4 any host 10.150.16.43
monitor capture po buffer size 10
monitor capture po limit pps 1000
PVW-CSW01#show monitor capture po buffer
buffer size (KB) : 10240
buffer used (KB) : 0
packets in buf : 0
packets dropped : 0
packets per sec : 0

PVW-CSW01#show monitor capture po buffer ?
brief brief display
detailed detailed disaply
dump for dump
| Output modifiers
<cr> <cr>

PVW-CSW01#show monitor capture po buffer br
PVW-CSW01#show monitor capture po buffer brief
----------------------------------------------------------------------------
# size timestamp source destination dscp protocol
----------------------------------------------------------------------------

[MHM Cisco World]

10.150.16.43 <<- this IP in VLAN16 ? if Yes try IP in VLAN18

MHM

[grapevine]

I shut and enabled the portchannel and I see this 

C:\Users\abc>ping 10.150.16.43 -t

Pinging 10.150.16.43 with 32 bytes of data:
Reply from 10.150.16.43: bytes=32 time=92ms TTL=58
Reply from 10.150.16.43: bytes=32 time=92ms TTL=58
Request timed out.
Reply from 10.150.16.43: bytes=32 time=94ms TTL=58
Reply from 10.150.16.43: bytes=32 time=92ms TTL=58
Reply from 10.150.16.43: bytes=32 time=92ms TTL=58
Reply from 10.150.16.43: bytes=32 time=92ms TTL=58
Request timed out.
Reply from 10.150.16.43: bytes=32 time=93ms TTL=58
Request timed out.

Ping statistics for 10.150.16.43:
Packets: Sent = 10, Received = 7, Lost = 3 (30% loss),
Approximate round trip times in milli-seconds:
Minimum = 92ms, Maximum = 94ms, Average = 92ms

[MHM Cisco World]

do capture and check 
if you see traffic in both way 
then this cons of using ON mode of PO, you need to force shut/no shut PO to make it work
MHM

[MHM Cisco World]

before run capture 
try shut/no shut PO since you run ON mode and one of stack SW is down 21 hrs ago

MHM

[grapevine · Accepted solution]

We powered off switch one by one in the stack, then noticed if the ESXI was reachable. the issue was that once the switch got added to the stack, the ESXI went unreachable

We figured out the issue. We issued the command spanning-tree portfast trunk on the portchannel and the issue resolved

We also had an issue with SFP that connected to ESXI2, we replaced it and it resolved.

Thank you all for the support

[MHM Cisco World]

I am 1000% sure issue appear again' 

With mode ON you need always to check PO.

Update me if it appear

MHM