Interface Up/Up on one side Down/Down on the other?

[tjalbrecht]

I have a 3750 that has a TW circuit connected to it.  I stopped incrementing packets input, but the link was up/up.  When I escalated to TW, they said their side was down/down.  I didn't think this was possible?

TW said they would bounce the interface.  I saw them shut it down, as my side went down/down.  They turned it back on, and my side went up/up.  Still no packets input from them, and they could not see my MAC.  The TW tech verified we are both 100/Full.  We changed that to Auto/Auto on both sides, and the link came up.  We then changed back to 100/Full and the link remained up.

I am thinking the TW tech had me change the speed/duplex to buy time, but I cannot put my finger on exactly what happened.  This is our primary link, and because of the way this circuit failed, EIGRP did not re-route us over the backup.  I tested failover just hours afterward by shutting this TW link down, and failover proceeded successfully.

Anyone experience this?  I think TW is not being open and honest, and this concerns me greatly.  The TW tech noted they had a 3400 connected directly into our switch, and they have a 7609 behind that.  Any thoughts by others in the field would be greatly appreciated.

Thank you,

T.J.

[Jon Marshall]

tjalbrecht wrote:
I have a 3750 that has a TW circuit connected to it.  I stopped incrementing packets input, but the link was up/up.  When I escalated to TW, they said their side was down/down.  I didn't think this was possible?
TW said they would bounce the interface.  I saw them shut it down, as my side went down/down.  They turned it back on, and my side went up/up.  Still no packets input from them, and they could not see my MAC.  The TW tech verified we are both 100/Full.  We changed that to Auto/Auto on both sides, and the link came up.  We then changed back to 100/Full and the link remained up.
I am thinking the TW tech had me change the speed/duplex to buy time, but I cannot put my finger on exactly what happened.  This is our primary link, and because of the way this circuit failed, EIGRP did not re-route us over the backup.  I tested failover just hours afterward by shutting this TW link down, and failover proceeded successfully.
Anyone experience this?  I think TW is not being open and honest, and this concerns me greatly.  The TW tech noted they had a 3400 connected directly into our switch, and they have a 7609 behind that.  Any thoughts by others in the field would be greatly appreciated.
Thank you,
T.J.

T.J

It really depends on whether the connection is literally a point to point or whether there are other switch(es) in between. If there are and there end goes down your end will still be up/up. This is one of the issues with ethernet.

If you want to remedy the situation in future you should take a look at IP SLA which allows you to monitor via ping the other end of the link and it fails you can then failover to the other link.

Edit - you would need the IPServices image on your 3750 to have IP SLA support.

Jon

[tjalbrecht]

Jon,

Thanks for the quick response.  I will look into IP SLA as I am not familiar with it.  The TW tech told me the 3400 switch is directly connected to our 3750, which does not make sense to me regarding the up/up and down/down descrepancy.  I have a call with them this afternoon to diagnose exactly what happened.

Thanks again,

T.J.

[Jon Marshall]

tjalbrecht wrote:
Jon,
Thanks for the quick response.  I will look into IP SLA as I am not familiar with it.  The TW tech told me the 3400 switch is directly connected to our 3750, which does not make sense to me regarding the up/up and down/down descrepancy.  I have a call with them this afternoon to diagnose exactly what happened.
Thanks again,
T.J.

T.J

When you do call them can you confirm with them where the IP address is that is your next-hop ie. is it on the 7609. If so it may be that the connection between the 3400 and the 7609 failed rather than the 3400 to you.

Jon