When changing the allowed VLANs per member interface the switch will give you the "(vlan mask is different)" message and suspend the port, which is logical as all the members must be compatible with each other, the correct action is as you said is to control the allowed vlans per the whole port channel and not individually.
On the other hand, it may not be as obvious as all that. I mean, I didnt add the trunk to the port channel port simply because it didnt occur to me to do it that way until about an hour into troubleshooting, I thought I had an L2 loop at first. So I made a mistake.
But going back to how different switches work, the 6513 CatOS switches are not like that. You can add the vlan to the individual trunk ports, as I tried to do on the 4506 yesterday.
And one would think that, having set the channel mode to desirable, that any change in a trunk port's allowable vlans would indeed cause a temporary suspension, but that once the other ports are matching (as you configure those to allow the additonal vlan), the channel mode would renegotiate to create the etherchannel once again.
I wonder if other manufacturer switches behave this way. I wonder if this is a Cisco thing or even just a Cisco native-IOS quirk.
We are pleased to announce availability of Beta software for 16.6.3. 16.6.3 will be the second rebuild on the 16.6 release train targeted towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are looking for early feedback from custome...