2960 Etherchannel Redundancy and Load-Balancing issue

[ali_777]

I have found 2 issues about port-channel when I used mode ON in Cisco 2960 Workstation switch.

The sample configuration in sw1 and sw2 are the same and is here:

Interface Port-Channel2
switchport mode trunk
!
Interface GigabitEthernet 1/1
switchport mode trunk
channel-group 2 mode ON
!
Interface GigabitEthernet 1/2
switchport mode trunk
channel-group 2 mode ON

 also port-channel summary shows

Group          Port-Channel         Protocol        Ports
--------------------------------------------------------
  2                Po2 (SU)                 ---                Gi1/1 (P)    G1/2 (S)

First: When One switch restart, then one of two port (that are bundled in ether-channel) change status to "Suspend" mode automatically after few minutes.

It means for some first minutes right after boot up it works normal (P) and everything is OK but after that the issue will occur (S) and need to change it back manually each time.

The Question here is How can I prevent it?

Second: It seems Load-Balancing works fine but I'm afraid that the redundancy work fine too in this mode. Actually I have seen that only half of all packets can be transmitted to destination by shutting down one of interface that are bundled in ether-channel. 

The Question here is how can I make redundancy to work fine too and auto stop to send to fault-link?

All suggestion about the issue are welcomed.

[Reza Sharifi]

Have you tried replacing the optic and the cable?

If none of this help, can you try a different port?

HTH

[ali_777]

Many Thnaks @paul driver and @james.berg for your suggestions.

Yes, as I explain bellow, connectivity can be cause of problem too but just does using Mode ON is enough for redundancy works (Without using any protocol and negotiation)?

---

@Reza Sharifi wrote:

Have you tried replacing the optic and the cable?

If none of this help, can you try a different port?

---

Dear @Reza Sharifi jan;

Thanks for your advise. Actually I suspected to it too. Because these are connected over a point to point bridge wireless link as show bellow:

  |----RADIO1----------------------RADIO2----|
SW1                                         SW2 
  |----RADIO3----------------------RADIO4----|

However P2P connectivity itself is steady good and stable. Maybe it cause the issue for etherchannel?

But another thing that I suspected is negotiation. I have found that using MODE ON making a static port-channel without using LACP or PAGP. That can cause no negotiation occurs between two switch.

---

---

 Thanks in advance.

[paul driver]

Hello

What is the 2960 switch connected to for aggregation?

[james.berg]

I believe your issue points to a cable issue of some sort
if fiber:
check your transceiver to make sure that all is in operational measures
sh interfaces transceiver detail

If copper
check & replace cables

Check for errors of any kind
sh int gi1/1
sh int gi1/2
sh int po2

review config
sh run int gi1/1
sh run int gi1/2
sh run int po2
show etherchannel 1 detail