** Description changed:

+ [ Impact ]
+ 
+  * Tests have shown that in very recent releases pollinate no more helps
+ but can slow down [1]
+ 
+  * At the same time hardware and the kernel learned plenty of tricks
+ which make it no more needed
+ 
+  * Therefore let us remove it from the default images in Resolute
+ onwards
+ 
+ [ Test Plan ]
+ 
+  * This is one of the trickier bits of this case. There are no
+    "proposed images" but the only place to really show it in non-action are
+    the final built images. But that would only be in effect once this update
+    is released to -updates. 
+    We could hand modify images, and for testing I have done that. But there
+    is no way to test "the real thing" earlier. So what can be done is done
+    already - is there nothing more?
+ 
+  * First of all this is gladly an important but not a complex change, we can
+    compare the dependencies of ubuntu-meta before/after the change and ensure
+    nothing but the intended change to the dependencies was done.
+ 
+  * And furthermore we can upgrade on an already installed system and
+    should see apt mentioning that pollinate is no more needed
+ 
+ [ Where problems could occur ]
+ 
+  * One could say "but then future images are different to past" which is not 
+    really a problem. pollinate really only does something on first boot. 
+    Therefore all the difference lies in the past. A system deployed a week 
ago 
+    and one next month after this change will still be identical in behavior 
as 
+    both their first boot lies in the past. And changing the future first 
boots 
+    is exactly what this is about.
+ 
+  * Someone could have hard depended or tied a service to pollinate, I could
+    not see a reason to do so, but this is one impact I could see that the
+    change has. But yet again this should be minimal - on new deployments
+    it won't be there and nothing in the pristine image depends on it. On
+    past deployments it will not vanish, apt will suggest to auto-remove
+    which one can do or not.
+ 
+  * Finally I was a bit concerned if some arcane art of image building
+    might be affected, but I checked with various people and got the ok
+    of the Ubuntu release team as well as the CPC team to go on with it.
+    That was in internal discussions, but to leave some breadcrumbs I also
+    asked them to leave an ok on the first seed change we did [2]
+ 
+ [ Other Info ]
+ 
+  * Sorry, I really wanted to do this before 26.04 was out but time was
+    not working out in my favor. Let us try to conclude it at least
+    before 26.04.1 please :-)
+ 
+ [1]: 
https://discourse.ubuntu.com/t/ubuntu-server-gazette-issue-16-does-pollinate-still-earn-its-place-on-every-ubuntu-boot-in-2026/84735
+ [2]: 
https://code.launchpad.net/~paelzer/ubuntu-seeds/+git/ubuntu/+merge/506815
+ 
+ --- original report ---
+ 
  Evaluations showed (a verbose version of that is soon to follow as a
  blog and will then be linked here - I still wait for a bunch of reviews)
  that what once was beneficial, and then often a no-op unless you have
  been in very special needs now became counter-productive.
  
  Drop pollinate from the default images in recent release please.

** Merge proposal linked:
   https://code.launchpad.net/~paelzer/ubuntu-seeds/+git/ubuntu/+merge/506815

** Merge proposal linked:
   https://code.launchpad.net/~paelzer/ubuntu-seeds/+git/ubuntu/+merge/507481

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2158262

Title:
  Please drop pollinate from the default installed set in ubuntu server

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/ubuntu-meta/+bug/2158262/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to