Hi Everyone, trying to update and it's failing, I've attached screenshots. I've tried migrating the LIFs manually, is it failing due to needing to define a failover group for the iSCSI LIFs?
#9.14.1P6 Failed to update to 9.15.1P4
near plinth
versed kernel
Some of the ports on node-01 appear to not be in the correct Broadcast Domain
This might be stopping the failover happening
silver fulcrum
I agree here. You should check the switches where a0a goes. Make sure the vlan is actually defined on all switches involved. If using Cisco and vpc make sure the vlan is on the peer link (some customers actually limit vlans there making troubleshooting a little more difficult)
When I see this it is usually the vlan was not defined on one switch.
Once you are sure, you can try
Net port reachability repair -node node-01 -port a0a-xxx
Net port reachability repair -node node-02 -port a0a-xxx
Give it a few seconds and look at the broadcast-domain show output.
Then you should be able to continue the upgrade
arctic ocean
to add to the explanation: the cluster needs to move the IPs away from the node that is about to reboot, so it tries to find one in the same L2 reachability domain on some other node. In your case it cannot find a port that is reachable, so it cannot move the LIF. You can either fix your switches' VLANs, or set the LIF's failover-policy to "disabled" which lets you upgrade but results in a multi-minute downtime of that IP address
silver fulcrum
Fix the networking
near plinth
I agree here. You should check the switches where a0a goes. Make sure the vlan i...
Reachability repair sends it to back default-1 after we added it to that broadcast domain
I am not a network guy do you know what I can tell the network engineer?
silver fulcrum
So the vlans are incorrect on the switch
silver fulcrum
Verify vlans 2084 and 2111 are on all involved switches.
Verify the port channels on all involved switches include vlans 2084 and 2112
Some switches seem to require putting the vlans in the port channel and the interface.
Either way make sure the vlans are identified everywhere there are supposed to be using vlan 2082/2083 as examples
near plinth
okay thanks hes sharing screen now I will ask
silver fulcrum
A “show vlan” on the switches should show those vlans.
If you know which interface
Show run int eth y/z
Will show the interface config and the vlans and the port channel
Show run int poZ
Will show the port channel config. It should look nearly identical to the port config.
If they are running nexus in vpc mode:
show vpc consistent-parameters global
This should show if both sides agree. If a vlan is missing it should let you know
near plinth
we added them and now everything says no reachability
they added them to the trunk
Sorry Lol
arctic ocean
im not sure how to ask him to add the VLAN port to the switch
"no L2 reachability" should be enough info for them to fix it 🤷♂️
silver fulcrum
Are they are Cisco switches using vPC?
near plinth
I've decided to get my CCNA after that whole expereince
near plinth
Are they are Cisco switches using vPC?
We added the vlans to the int's but not port channel and then once we added to port channel it worked like a charm
Thank you everyone you guys are the best
silver fulcrum
You are welcome!
versed kernel
i would also rename those 2 broadcast domains on Node 1 to keep it consistent with node 2
silver fulcrum
Broadcast domains are cluster resources. Once the ports are together, there will likely be a “default-X” broadcast domain left over that is empty
arctic ocean
yeah, either wait & hope for ONTAP to merge them (the reachability repair command might help there) or just merge the broadcast domains manually. Then you're set