I have been facing 2 issues since we went live with our new infra last week.
Issue#1
On the single EX2028P, I am using Ex0/1 and Ex0/2 for uplinks to EX2028P cluster which are connected to the core switches . I have set them manually to Trunk ports. But whenever switch is restarted, it auto reverts to Hybrid with PVID 1. I have to then run to the switch and connect mgmt port and change the port type. Also the port speed for all Ex ports revert to 10gbps, which I have already set to 1gbps as we are using 1gbps SFP modules.
Issue#2
We have uplinks from single EX2028P to the EX2028P cluster for redundancy. Twice I have found that if the cluster got rebooted, the single EX2028P would not take a DHCP lease from the Fortigate gateway. The 2nd time was just an hour ago as of writing this.
All the switches are or 3.0.1-r4. Is this some kind of bug. I have tried to boot default and reconfigure the switches again but no use. This is a production site and it’s very difficult to get downtime for short time also.
Regarding issue #1. Please make sure to run CLI command ‘write startup-config’ or click the Save button on the Web GUI as shown after you finish the configuration change.
Regarding issue #2. I am not clear with the problem description. Do you mean the single EX2028P did not release its DHCP IP address obtained before from the Fortigate gateway when the trunk ports are down? The IP address is associated with the VLAN interface status. I suspect the VLAN in which the trunk ports are member of is ‘Up’ because other ports in the VLAN are still up.
I am using CLI mostly so running copy running-config startup-config command always whenever I am doing any changes. And also clicking on Save from the GUI if changes done from GUI. But only these 2 specific things, the speed and port type only on the EX ports revert.
Lets consider this example. The single EX2028P had leased the IP 172.41.190.171 when the interface was up and everything working properly as expected. Suddenly, the EX2028P cluster was rebooted due to power failure in the network cabinet. Once the cluster was back up, the single EX2028P should come back up and go online. But this did not happen. The interface itself was up and the cluster was also online. Finally after few minutes I just rebooted and it came back online immediately.
Either ‘copy running-config startup-config’ or ‘write startup-config’ should save the configuration in flash device. We are aware of any configuration lost after reboot, and I am unable to replicate the symptom just now. Can you please send me the output of ‘show running-config’?
I would need a remote session to investigate further on #2 issue.
I have reviewed your running-config. The reason your EX ports revert to hybrid from trunk is because they are part of a port-channel. When a port participates in a port channel, its configuration is defaulted. Its behavior adheres to the configuration of port channel. In this case you need to make the port-channel a trunk as shown:
config terminal
interface port-channel 10
switchport mode trunk
