Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Lennart Poettering on opentmpfiles: it's just baaaad
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5, 6, 7  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 46272
Location: 56N 3W

PostPosted: Sun Dec 29, 2019 11:22 pm    Post subject: Reply with quote

saellaven,

saellaven wrote:
Why does lm-sensors depend on openrc-0.36 or newer when my 0.17 works fine, unmodified? Because of the politics of trying to force people into something they don't want, which ties back into WilliamH and his crusade to cripple openrc to make sure it never gets ahead of systemd in functionality. What do all the systemd proponents say about competition? Write something better... and how does that happen when things are intentionally hamstrung to not make systemd look inferior?


You are mistaken.
Round about openrc-0.18, some functionality was moved out of openrc into a separate package.
It didn't last very long and that functionality was merged back into openrc again.
Portage syntax does not permit version ranges to be specified, so ebuild writers need to use >openrc-0.x or >=openrc-0.y and not permit the lower range that still works.
>openrc-0.12 would fail for some openrc versions.

In my opinion, there were some questionable openrc design decisions made around openrc-0.18, some of which were fixed later.
I stuck with 0.17 because it still works and I don't have time to review later versions to fix problems I don't have.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Naib
Watchman
Watchman


Joined: 21 May 2004
Posts: 5847
Location: Removed by Neddy

PostPosted: Sun Dec 29, 2019 11:31 pm    Post subject: Reply with quote

GDH-gentoo wrote:
Naib wrote:
GDH-gentoo wrote:
Naib wrote:
The need for an initramfs to manage /usr on a separate partition [...] for some reason it also impacted OpenRC
I keep reading this, and if it means that OpenRC does not support booting a system with /usr on a separate filesystem and without an initramfs, then it is not true, at least for current stable branch OpenRC, as long as there is an implementation of md5sum in the rootfs.
The issue is actually udev. so if you use udev (or even eudev) then you will need to have an initramfs because aspects of udev are stored in /usr/ [...]
(I think I posted about this before)

Not true either, at least for current stable branch eudev. udevd is in /sbin, udevadm is in /bin, the shared libudev is in /lib or /lib64, the rules themselves are in /lib/udev/rules.d, and the ones installed by the eudev package itself either call udev builtins or programs in /lib/udev. How do I know it works? Because I have a system with eudev, separate /usr and no initramfs, and it boots with both sysvinit + OpenRC and s6 + s6-rc + s6-linux init.

Naib wrote:
it doesn't' matter which init system you use (systemd, OpenRC, r6, busybox...) *IF* at any point an init service requires something from a path that isn't mounted you will not be able to launch that service.
This (poor packaging practices) is the real issue, but the offenders would be the corresponding packages, not OpenRC or eudev.
Anecdotal evidence doesn't provide 100% coverage. It works for me as well, but I don't have a /usr partition so I fail to see how your anecdotal evidence refutes my clear statement of what can limit /usr on another partition.

it is quite simple... to boot your system you need the libraries/applications available. It doesn't matter if it is on a separate partition, a slow storage device (usb for instance), encrypted drive, NFS. to boot the system the prerequisites are needed FACT. the larger binary distro's try to put EVERYTHING in to provide all combinations. Gentoo offers the end-user the ability to tailor to their needs. Result? a higher chance these use-cases that systemd box themselves into are likely to be of lesser concern on a gentoo system

take this:

egrep 'usb-db|pci-db|FROM_DATABASE|/usr' /*/udev/rules.d/*


Simple command and on my system I have udev rules that would launch applications in /usr/sbin HERPDERP these would fail on a non-mounted /usr. For me these are just sound so are trivial for me and less of a concern as I have a flat partition structure


can it work? sure, if you do not have anything critical for boot residing in /usr. does it universally work? apparently not because of poor architectural decisions baked into udev (which could be trivially worked around)
_________________
The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king
Back to top
View user's profile Send private message
GDH-gentoo
Guru
Guru


Joined: 20 Jul 2019
Posts: 482
Location: South America

PostPosted: Mon Dec 30, 2019 2:11 am    Post subject: Reply with quote

Naib wrote:
Anecdotal evidence doesn't provide 100% coverage. [...]

[...] to boot the system the prerequisites are needed FACT. [...] Gentoo offers the end-user the ability to tailor to their needs. Result? a higher chance these use-cases that systemd box themselves into are likely to be of lesser concern on a gentoo system

[...]

Simple command and on my system I have udev rules that would launch applications in /usr/sbin HERPDERP these would fail on a non-mounted /usr. [...]


can it work? sure, if you do not have anything critical for boot residing in /usr. does it universally work? apparently not [...]
We are not in disagreement here. I said:
Quote:
It is true that separate /usr with no initramfs is not a supported setup on Gentoo. Which doesn't necessarily mean that it won't work, only that whoever wants this setup is on his/her own. It is true that depending on which packages are installed, it might not work indeed. One common packaging error that causes this being udev rules that call executables in /usr [1].
But this started with what I perceived to be claims that, to the best of my knowledge, are misconceptions, and seem to be be repeated quite often around here. To summarize:
  • Does current stable branch OpenRC force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
  • Does current stable branch eudev (the package, not udev rules that other packages might install) force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
If both were true, it would be impossible to have a Gentoo system with eudev, OpenRC, split /usr and no initramfs. I know of at least one such system, therefore, they are not true.
  • Can other packages make a split /usr and no initramfs setup fail? Probably yes. But it would be those packages' fault, not OpenRC's or eudev's
Back to top
View user's profile Send private message
saellaven
Guru
Guru


Joined: 23 Jul 2006
Posts: 569

PostPosted: Mon Dec 30, 2019 2:34 am    Post subject: Reply with quote

NeddySeagoon wrote:
saellaven,

saellaven wrote:
Why does lm-sensors depend on openrc-0.36 or newer when my 0.17 works fine, unmodified? Because of the politics of trying to force people into something they don't want, which ties back into WilliamH and his crusade to cripple openrc to make sure it never gets ahead of systemd in functionality. What do all the systemd proponents say about competition? Write something better... and how does that happen when things are intentionally hamstrung to not make systemd look inferior?


You are mistaken.
Round about openrc-0.18, some functionality was moved out of openrc into a separate package.
It didn't last very long and that functionality was merged back into openrc again.
Portage syntax does not permit version ranges to be specified, so ebuild writers need to use >openrc-0.x or >=openrc-0.y and not permit the lower range that still works.
>openrc-0.12 would fail for some openrc versions.

In my opinion, there were some questionable openrc design decisions made around openrc-0.18, some of which were fixed later.
I stuck with 0.17 because it still works and I don't have time to review later versions to fix problems I don't have.


it's been, what, about 5 years? The memory is a little rusty about the details at this point, but yes, that openrc-0.18 functionality nonsense was exactly what caused me to stay at 0.17 too. I considered openrc to be potentially damaged intentionally for political reasons going forward and couldn't justify using new versions without taking the time to code review (also why I abandoned gentoo-sources for vanilla-sources after TomWij became a maintainer of gentoo-sources and an ardent defender of WilliamH's political moves, despite considering becoming a gentoo dev focusing on the kernel around the same time myself).

Way too many of this crowd DO before thinking and after blowing stuff up, will revert, not caring for all the damage caused in the meantime. The same people are still doing the same things today (see the recent python-2.7 disruption caused by an overeager beaver, GLEP 81 waffling, etc).
_________________
Ryzen 3700X, Asus Prime X570-Pro, 32 GB DDR4 3200, GeForce 1050 Ti
openrc-0.17, ~vanilla-sources, ~nvidia-drivers, gcc-9.3.0
Back to top
View user's profile Send private message
saellaven
Guru
Guru


Joined: 23 Jul 2006
Posts: 569

PostPosted: Mon Dec 30, 2019 2:37 am    Post subject: Reply with quote

GDH-gentoo wrote:
Naib wrote:
Anecdotal evidence doesn't provide 100% coverage. [...]

[...] to boot the system the prerequisites are needed FACT. [...] Gentoo offers the end-user the ability to tailor to their needs. Result? a higher chance these use-cases that systemd box themselves into are likely to be of lesser concern on a gentoo system

[...]

Simple command and on my system I have udev rules that would launch applications in /usr/sbin HERPDERP these would fail on a non-mounted /usr. [...]


can it work? sure, if you do not have anything critical for boot residing in /usr. does it universally work? apparently not [...]
We are not in disagreement here. I said:
Quote:
It is true that separate /usr with no initramfs is not a supported setup on Gentoo. Which doesn't necessarily mean that it won't work, only that whoever wants this setup is on his/her own. It is true that depending on which packages are installed, it might not work indeed. One common packaging error that causes this being udev rules that call executables in /usr [1].
But this started with what I perceived to be claims that, to the best of my knowledge, are misconceptions, and seem to be be repeated quite often around here. To summarize:
  • Does current stable branch OpenRC force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
  • Does current stable branch eudev (the package, not udev rules that other packages might install) force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
If both were true, it would be impossible to have a Gentoo system with eudev, OpenRC, split /usr and no initramfs. I know of at least one such system, therefore, they are not true.
  • Can other packages make a split /usr and no initramfs setup fail? Probably yes. But it would be those packages' fault, not OpenRC's or eudev's


What you're missing, is there are supported patches that DO ensure you can boot without an initramfs but those patches were specifically excluded to ensure that openrc doesn't appear more robust than systemd... because one of the (mis)leading arguments for systemd is that it is technically superior to everything else.
_________________
Ryzen 3700X, Asus Prime X570-Pro, 32 GB DDR4 3200, GeForce 1050 Ti
openrc-0.17, ~vanilla-sources, ~nvidia-drivers, gcc-9.3.0
Back to top
View user's profile Send private message
Naib
Watchman
Watchman


Joined: 21 May 2004
Posts: 5847
Location: Removed by Neddy

PostPosted: Mon Dec 30, 2019 8:25 am    Post subject: Reply with quote

saellaven wrote:
GDH-gentoo wrote:
Naib wrote:
Anecdotal evidence doesn't provide 100% coverage. [...]

[...] to boot the system the prerequisites are needed FACT. [...] Gentoo offers the end-user the ability to tailor to their needs. Result? a higher chance these use-cases that systemd box themselves into are likely to be of lesser concern on a gentoo system

[...]

Simple command and on my system I have udev rules that would launch applications in /usr/sbin HERPDERP these would fail on a non-mounted /usr. [...]


can it work? sure, if you do not have anything critical for boot residing in /usr. does it universally work? apparently not [...]
We are not in disagreement here. I said:
Quote:
It is true that separate /usr with no initramfs is not a supported setup on Gentoo. Which doesn't necessarily mean that it won't work, only that whoever wants this setup is on his/her own. It is true that depending on which packages are installed, it might not work indeed. One common packaging error that causes this being udev rules that call executables in /usr [1].
But this started with what I perceived to be claims that, to the best of my knowledge, are misconceptions, and seem to be be repeated quite often around here. To summarize:
  • Does current stable branch OpenRC force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
  • Does current stable branch eudev (the package, not udev rules that other packages might install) force the use of an initramfs to pre-mount /usr if it is a separate filesystem? No.
If both were true, it would be impossible to have a Gentoo system with eudev, OpenRC, split /usr and no initramfs. I know of at least one such system, therefore, they are not true.
  • Can other packages make a split /usr and no initramfs setup fail? Probably yes. But it would be those packages' fault, not OpenRC's or eudev's


What you're missing, is there are supported patches that DO ensure you can boot without an initramfs but those patches were specifically excluded to ensure that openrc doesn't appear more robust than systemd... because one of the (mis)leading arguments for systemd is that it is technically superior to everything else.
and thus if a patch to those parts results in a split /use being bootable then aspects of those applications are associated with the problem
_________________
The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 46272
Location: 56N 3W

PostPosted: Mon Dec 30, 2019 9:00 am    Post subject: Reply with quote

saellaven,

saellaven wrote:
Way too many of this crowd DO before thinking and after blowing stuff up, will revert, not caring for all the damage caused in the meantime. The same people are still doing the same things today (see the recent python-2.7 disruption caused by an overeager beaver, GLEP 81 waffling, etc).


Thats a universal problem, not just systemd and its proponents. Hence
Jack Bergman wrote:
There's never enough time to do it right, but there's always enough time to do it over.


The UNIX way of do one thing and do it well and making complex systems out of lots of little pieces means that the entire problem at hand (one of the little pieces) is small enough to be kept is the designers head.
Traditionally *nix has been saved from the mess that results when little or no system design is performed by that philosophy.
Unfortunately, until they have been on a couple of failed projects, its not something that most people learn.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7447

PostPosted: Mon Dec 30, 2019 12:41 pm    Post subject: Reply with quote

Naib wrote:
Anecdotal evidence doesn't provide 100% coverage. It works for me as well, but I don't have a /usr partition so I fail to see how your anecdotal evidence refutes my clear statement of what can limit /usr on another partition.

Here's a real example : https://forums.gentoo.org/viewtopic-t-1076204.html
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 4111
Location: Illinois, USA

PostPosted: Sun Jul 26, 2020 3:27 am    Post subject: Reply with quote

NeddySeagoon wrote:
Traditionally *nix has been saved from the mess that results when little or no system design is performed by that philosophy.
Unfortunately, until they have been on a couple of failed projects, its not something that most people learn.

I recommend the book "Death March" by Edward Yourdon. It's about software projects going awry, not WW II.
Instead of a commercial link I'll just say "Check your favorite bookstore, brick and mortar or on-line".
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18556

PostPosted: Sun Jul 26, 2020 7:12 pm    Post subject: Reply with quote

Split off crocket's corner.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3, 4, 5, 6, 7
Page 7 of 7

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum