09-10-2019 11:02 PM
The VM are running on Nutanix. Each node have 2 NIC. Nic 1 to SW 1 & NIC 2 to SW 2. You may refer the diagram.
As mention by them, they did test by ping the VM and shutdown the host. So the host are connect to both SW 1 & 2.
And then customer mention need the uplink fast to be configure. Otherwise the packets will drop more then 20. You may refer the attch the test.
09-11-2019 06:43 AM
I aplogize but I do not think I have enough to go on to provide you with a proper analysis. If I could receive a show tech form each switch this could show me what is going on. i would prefer this to be gathered after they attempt the test they are doing? I also, have some follow up questions:
09-12-2019 09:00 AM - edited 09-12-2019 10:42 AM
Thanks for this information.
Looking through the configuration I do not see any VLAG configuration. Is the top portion of the show run missing? and the ports don't seem to be configured as a lag which is necessary to create a VLAG instance. Would it be possible to send me a "show tech" from each switch in separate files. I would recommend utilizing some type of logging within your terminal because a show tech can be very large.
Also, can you let me know what port#s you are using to connect up to the servers? Based on the configuration it looks like 1, 3 and 5 but I wanted to make sure.
09-13-2019 09:01 AM
Let me brief actually what is the early plan:-
1) Customer purchase Nutanix with 3 nodes (each node with 2 NIC)
2) And mistakenly purchase Lenovo NE1032 x 2 unit.
The early plan was, connect all nodes NIC to direct to their customer so call core sw which is i mention Huawei (single unit).
But once customer get to know the have the Lenovo sw 2 unit, they would like to use as stacking.
After further studies, was found the Lenovo NE1032 does not support stacking. It only able work as VPC which is term for Cisco. But for VPC to both Lenovo sw said need to do VLAG.
There is came the VLAG part.
So with all 6 fiber connetion from Nutanix are connected to Lenovo Sw 1 (3 fiber) & sw 2 (3 fiber). And from customer Huawei, they only have 2 port to give to us. Which mean cureent we connect 1 port to sw1 & another port to sw2. And customer site they did a LACP.
SO at Lenovo port, connection need to be created between Lenovo sw1 & 2. Thats the part we currently working on it. With port 27, port 31 peer link. And port 32 back to Huawei.
And as for port 1,3, & 5, yes its connect back to Nutanix nodes.
I'm unable to get the config on top, as away.
So base on this, from stacking. any odea that i could work out? Agin customer sw only have 1 unit off Huawei. So no redunduncy.
09-16-2019 07:22 AM
It would be really helpful if you could get me the rest of the configuration. Considering stacking is not an option with these switches we need to make sure VLAG is configured correctly or you will see some issues. From the looks of it the connections to the switch and the configuration on the servers and the Huawei switch seems correct.
The recommendation I would make is to make sure you have a minimum two port lag in between the NE switches as your ISL for VLAG.
09-16-2019 08:35 AM
Again your are doit good and helping me...thanks again
As mention I'm still away from office and customer. But i did have something on VLAG.
VLAG info..probably u may look on this ..