On 23/11/2024 at 14:32, Holger Wansing wrote:
Pascal Hambourg <pas...@plouf.fr.eu.org> wrote (Sat, 23 Nov 2024 14:17:31
+0100):
On 23/11/2024 at 14:03, Holger Wansing wrote:
but the main issue (a heavy swap partition eats up most disk space) is
now solved with the server recipe, so I think it's ok to close this report.
This bug report could even have been closed in 2020 after the
introduction of partman-auto/cap-ram. I just wanted to point out that
the issue is solved with all built-in recipes, not only the server one.
Things have changed here recently:
from 2020 until September of this year you were right: it worked with any
recipe (because of cap-ram -> 1G).
However, in October we pushed in an overhaul of recipes, to improve
calculation of partition sizes, and in-line with this, the cap-ram
functionality was disabled by default; we needed this, to be able to get
swap sizes bigger than 1G in auto-partitioning (and with this, to make
hibernation work again on many machines).
So, *now* it's only working with the server recipe on such machines.
I beg to disagree. The new PRIORITY values defined in other recipes
(atomic, home, multi, small_disk) were designed to limit the swap size
to approximately 5% of disk size. So even with cap-ram disabled, the
swap cannot eat most disk space with any recipe.
