On 2021-11-11 11:59, Ulrich Mueller wrote:
We could:

- Open some part of the range between 500 and 1000. For example,
   500..799, which would leave 200 IDs for dynamic allocation.

- Open part of the range 60001..65533. Not sure if all software will be
   happy with that.

- Admit that the concept of static allocation has failed, and return to
   dynamic allocation.

Only the third option is really possible.

The first option (500-1000) would be technically possible but would clash with knowledge people gained in the past and would violate LPIC (=making Gentoo even more special and unusable for companies relying on certifications).

In addition, it would just delay the problem we currently have and not solve/address it.

Allowing ranges 60001+ is technically not an option. Expect that daemons using IDs >1000 will run into problems. Expect security problems because known system user range is hardcoded in many places so 60001+ is unexpected. This will really make Gentoo 'unique' in a really bad way and will break with everything which is/was being taught/documented in the world.

Let's face it: The idea of static ID allocation didn't scale. Let's stop this experiment before it is too late. Like you know, I always ask why someone is proposing a change, i.e. asking for the motivation. The main driver behind static IDs was that when you are maintaining multiple systems, that if IDs are identical, it will make life a little bit easier because you could copy files from service A on system 1 to service A on system 2 without the need of adjusting permission afterwards. But is this really a problem? From my POV it isn't:

1) If this really was bothering you, you already had a solution in place. Keep in mind: Most setups don't just consist of Gentoo/Debian/RHEL-only... you usually have a mix of setups so you need a solution which works everywhere so you don't need that 'feature' Gentoo offered (not to mention that you probably have something like AD in place which will make things like that very easy).

2) Pay attention to the way how you do stuff today. You will not create systems manually anymore (and if you do, you would just clone so there isn't even a need for this). You will automate this in scripts and use tools like Ansible, Salt, Chef, Puppet.... and of course, Dockers (which is basically a script) and like mentioned, AD.

From my POV I cannot imagine a single reason why we should stick to this idea and invest more time into it with the risk of making Gentoo more unique causing more _severe_ problems in future.

Anyone who wants to keep this around and wants to extend UID ranges instead should answer the following questions:

1) How are you going to solve the mentioned problems?

2) Why do you believe this feature is worth all the trouble?

3) At the moment we can stop. But once we start altering systems to mark additional ranges for system users there is _no_ easy way back anymore. Any blow up will probably require user to reinstall their entire system...


--
Regards,
Thomas Deutschmann / Gentoo Linux Developer
fpr: C4DD 695F A713 8F24 2AA1 5638 5849 7EE5 1D5D 74A5

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

Reply via email to