Has anyone experienced issues when moving between awsx and awsx/classic? Specifically around EKS trying to recreate deployments (instead of updating). This is with just a node_module update and no other changes. Just wanted to see if there was any tips or tricks to getting this stuff to leave unchanged deployments alone.

There are significant differences between

awsx/classic

and

awsx

, so I’m not surprised that you are seeing proposed changes. There is no workaround of which I am aware, sorry.

I thought the whole point of moving to /classic was to allow for no changes? if I don't use classic I have to rewrite pretty much everything lol. It's not a super pleasant experience.

Moving to

awsx/classic

allows you to “opt out” of the changes, as long as you don’t upgrade out of it. Unfortunately, getting away from classic means rewriting stuff---sorry about that.

You're upgrading from 0.4.1 to something newer, and changing all the imports from awsx.* to awsx.classic.*? And everything except EKS deploymets are doing the right things?

I took his question the opposite way, but maybe I misunderstood? I thought @gray-chef-51307 was already on

awsx/classic

and was moving to

awsx

.

You don't have any changes in node_modules when changing API within the one package...

No, @little-cartoon-10569 is correct.

Can you paste the detailed diffs of the unexpectedly-changing resources? Slacks' Text Snippet tool is good for sharing fixed-width formatted text.

It's quite a bit. Let me find my last pulumi up and post them both

All the deployment replaces are due to changed specs, which is correct. Are the specs changing unexpectedly? Can you view the detailed diff, rather than the summary, and inspect the spec changes? If they're different for a "bad" reason and causing inappropriate replacements, then I'm sure Scott and the Pulumi team would like to know.