Paulow21978
Contributor
Contributor

MPIO port binding issue

Jump to solution

Hi All,

I am having a problem with setting up MPIO round robin on my Dell R630 servers linked to a new Dell EqualLogic PS4100. I have done this before many times in previous incarnations of VMware and Dell servers so its a bit of a weird one. I recently updated to ESXi 6.7 as well.

I have 6 servers access the San and this issue is affecting all of them. I have done the usual of adding 2 VMKernel ports to a vswitch and overriding the failover, dropping one nic out of each port as required then adding the port binding in to the iscsi HBA. I have 6 san luns on the new San and weirdly, I can see 6 paths to the SAN (1 for each SAN lun) and the paths vary across the two ip addresses I gave the two vmkernel ports. I have also tested using vmkping from another esxi host and it can ping both the vmkernel ports and the two equallogic ip addresses.

In the port binding panel it says that one port is "Last active" and the other is "not used". I am not using any vlans and the switches they are going through are a coupe of stackable Dell switches which are only used for iSCSI. I have checked the ip's and ruled them out as they are pingable. Everything is set to jumbo frames. I did take one of the ports out of the port group and rescanned and the net result was no datastores from the san. I then readded the one I removed and removed the one I had left in previously and rescanned then the datasotres came back. I have checked the equallogic and its been locked down to the initiator and both nics are enabled and configured correctly.

Wierdly, on the vmkernel port group settings one nic has some CDP information and the other does not. I haven't used CDP before so I don't kno if this would cause some weird issue.

If anyone has any ideas it would be much appreciated. Quite why one of the ports in the network port binding will not return any datastores in the paths list BUT is pingable using vmkping from another ESXi host is a bit of a mystery but I suspect there is something in there that might be the cause of the issue.

if you have any ideas please help!

many thanks

Paul

0 Kudos
1 Solution

Accepted Solutions
Paulow21978
Contributor
Contributor

Well, I eventually got to the bottom of this. We are using Dell N series switches (10GB in a stack of two units. One of them had disconnected from the stack master and was not passing traffic between the stacking cable. Makes sense!

View solution in original post

0 Kudos
1 Reply
Paulow21978
Contributor
Contributor

Well, I eventually got to the bottom of this. We are using Dell N series switches (10GB in a stack of two units. One of them had disconnected from the stack master and was not passing traffic between the stacking cable. Makes sense!

0 Kudos