I am experiencing connectivity issues in my UCS environment. An ESXi host with vnic0 connected to I/O Module 1 (2104) and vnic1 to I/O Module 2 is having connectivity issues to rest of network. When I remove vnic1 from the host, IP connectivity is restored. Attempting to use only vnic1 and not connecting vnic0 to the vswitch cause loss of IP connectivity to the hosts. I have tried Active/Standby, Active/Active configurations on the vswitch and same results when I connect vnic1.
Can someone please tell me what I should be looking for in the UCS environment or how I can fix this problem please.
Ok that's good.
Can you follow the below action plan and let us know how it goes
1. Modify vnic 1 to go through the same path as vnic 0, so both of them must be going out through the same IOM and FI. Now remove vnic 0, do you still have connectivity ?
If yes, proceed to step 2
If no, we'd need to look deeper into the properties/config of vnic 1
Step 2. We could be staring at possibly a faulty port on the adapter or IOM or chassis back plane port, let us now proceed with eliminating each one.
Can you insert the blade in question with a vacant slot in the same or another chassis?
If no slots are vacant, can we swap with a blade in another slot, is this a possibility ?
In this step we are trying to isolate if the issue is with the port on the adapter.
After the blade is moved into a different slot, if the issue goes away (make sure vnic0 and vnic 1 are not going through the same IOM/FI, which was done in earlier activity) then the IOM back plane port or the chassis back plane port could be culprit here. Go to step 3.
If the issue remains, (physically reseat the adapter just to make sure that the adapter is seated correctly) the port on the adapter is bad and the adapter needs to be replaced. (you could further verify this by swapping a known good adapter, if need be.)
Step 3. Place the blade back to the slot where the issue existed, swap the IOM's. If the issue remains, the chassis back plane port is the culprit and the chassis needs to be replaced (the possibility of the chassis back plane port being bad is slim, but we've seen these corner cases too)
If the issue goes away on vnic 1, the IOM's back plane port is at fault and the IOM needs to be replaced.
Let us know how this goes.
I will give try out the steps you provided. New development, not sure of what was going on, I disabled and enabled the NIC1 in UCS Manager and now I can have both NICS connected to the esxi host and not loose connectivity. Tried the same on another server and it did not make a difference. Weird right