• Zotero can't run on Jessie, so can't be moved.
• Idem for Apertium, even though there's a WIP to package it for Jessie.
• CP is on a single machine because it can become CPU-hungry, so we don't want it to bring down other services in Beta when/if that happens. Also, we use it heavily for testing real-world scenarios. IMHO, better to leave this one as-is.

Hence, invalid? :)

• Citoid: Created T142150, blocked by T107302
• Apertium: Created T142152, blocked by T107306

I'll leave CP alone.

I'll note here, however, that "you" (anyone in charge of any service) should have an explicit step in their upgrade path of upgrading the service in Beta Cluster *before* upgrading in production. This has routinely been missing (at least explicitly in any task I've seen, whether or not it actually gets done).

In T142100#2524532, @greg wrote:

• Citoid: Created T142150, blocked by T107302
• Apertium: Created T142152, blocked by T107306

I'll leave CP alone.

Awesome, thnx @greg !

I'll note here, however, that "you" (anyone in charge of any service) should have an explicit step in their upgrade path of upgrading the service in Beta Cluster *before* upgrading in production. This has routinely been missing (at least explicitly in any task I've seen, whether or not it actually gets done).

Agreed. I do that usually without explicitly mentioning it. But, more to your point, I'm confused as how this relates to the task of moving services to the same host in Beta; the process of deploying in Beta is still manual, regardless of the target. To that, I have talked a long time ago to @hashar about getting this process automated and I would love to see it happen.