You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think it's ok having a distribution in active LTS be used so sticking with bookworm would be viable (Part of me wonders if explicitly sticking with one in the LTS phase is a good idea for stability!) but if we don't have a policy on these upgrades we should probably look at getting one defined.
But given where we are at the moment, I'm personally ok with switching the default - do we need to announce such changes anywhere first?
I do wonder whether for Debian that June date isn't so much of a notable event since major architectures (arm and x64) should still receive public (without a subscription) support from a security perspective for much longer from the LTS team (so on that basis I wouldn't necessarily recommend that we stop publishing bookworm images in June)
So basically we should also consider the policy for stopping old images as well as when to switch to new ones - worst case we continue publishing them until the OS goes out of public Long Term Support - which is likely when Dockerhub's official images would stop supporting it too.
This issue was only intended to be about setting the default image to trixie, not about stopping distributing bookworm based images when they are explicitly selected.
Nevertheless it would definitely be helpful to put down some guidelines for Alpine and Debian about introduction and retirement based on upstream lifecycle dates. I don't believe that is described anywhere in this repo.
I am going to close this issue, as it needs a more granular approach depending on the version of Node.js involved, and this would be part of a more general release cycle determination.
Activity
MikeMcC399 commentedon Jan 30, 2026
If this makes sense and is welcome I'll submit a PR to implement.
sxa commentedon Jan 30, 2026
I think it's ok having a distribution in active LTS be used so sticking with bookworm would be viable (Part of me wonders if explicitly sticking with one in the LTS phase is a good idea for stability!) but if we don't have a policy on these upgrades we should probably look at getting one defined.
But given where we are at the moment, I'm personally ok with switching the default - do we need to announce such changes anywhere first?
MikeMcC399 commentedon Jan 30, 2026
@sxa
That does seem to be a generally missing document that describes what's included, or not, and how changes are decided on.
Regarding the update itself, perhaps just start with adding some text to the README?
bookwormis the current default OSbookwormtotrixieas defaultInstead of making the change now with no advanced warning, first choose a planning date and publish it.
The EOL date for bookworm might be appropriate (2026-06-10, or the release of Node.js 26 (2026-04-22), for instance.
sxa commentedon Jan 30, 2026
I do wonder whether for Debian that June date isn't so much of a notable event since major architectures (arm and x64) should still receive public (without a subscription) support from a security perspective for much longer from the LTS team (so on that basis I wouldn't necessarily recommend that we stop publishing bookworm images in June)
So basically we should also consider the policy for stopping old images as well as when to switch to new ones - worst case we continue publishing them until the OS goes out of public Long Term Support - which is likely when Dockerhub's official images would stop supporting it too.
MikeMcC399 commentedon Jan 30, 2026
@sxa
This issue was only intended to be about setting the
defaultimage totrixie, not about stopping distributingbookwormbased images when they are explicitly selected.Nevertheless it would definitely be helpful to put down some guidelines for Alpine and Debian about introduction and retirement based on upstream lifecycle dates. I don't believe that is described anywhere in this repo.
MikeMcC399 commentedon Feb 10, 2026
I am going to close this issue, as it needs a more granular approach depending on the version of Node.js involved, and this would be part of a more general release cycle determination.