Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Do you use an initramfs?
View unanswered posts
View posts from last 24 hours

Goto page 1, 2, 3, 4  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  

Do you use an initramfs?
Yes, most distros do so I guess it's the right move
2%
 2%  [ 2 ]
Yes, I use genkernel so it's a no-brainer
7%
 7%  [ 7 ]
Yes, I've got a system setup that requires one
26%
 26%  [ 24 ]
Yes (Other - please explain)
5%
 5%  [ 5 ]
No, I do my own kernel so it's just more work
39%
 39%  [ 36 ]
No, it's just something else to go wrong
10%
 10%  [ 10 ]
No (Other - please explain)
8%
 8%  [ 8 ]
What's an initramfs?
0%
 0%  [ 0 ]
Total Votes : 92

Author Message
Havin_it
Veteran
Veteran


Joined: 17 Jul 2005
Posts: 1240
Location: Edinburgh, UK

PostPosted: Thu Oct 31, 2013 1:37 pm    Post subject: Do you use an initramfs? Reply with quote

Just curious. We're one of the few distros where using an initrd/initramfs isn't pretty much mandated, but I know a lot of us (most?) choose to use one anyway, so I wondered what people's reasons are for using one (or not).

Personally I started out not using genkernel, and with a simple setup, so never had a need for one. But I do wonder what I'm missing sometimes.
Back to top
View user's profile Send private message
GFCCAE6xF
Apprentice
Apprentice


Joined: 06 Aug 2012
Posts: 295

PostPosted: Thu Oct 31, 2013 1:41 pm    Post subject: Reply with quote

No (Other - please explain):
My setup (well at least the one I am writing this from) does not require one to function, if it did I would happily use one like I do elsewhere :)
Back to top
View user's profile Send private message
John R. Graham
Administrator
Administrator


Joined: 08 Mar 2005
Posts: 10587
Location: Somewhere over Atlanta, Georgia

PostPosted: Thu Oct 31, 2013 1:48 pm    Post subject: Reply with quote

No (Other - please explain): I don't need one right but I'm not philosophically opposed to one. An initramfs is only a little extra work which is convenient to avoid for now and, more to the point, has no effect whatsoever on the eventual runtime. Certain configurations that I intend to experiment with in the future (e.g., encrypted root) will require one so I'll likely have systems both with and without at some point. No big deal—and no need for a fuss—in my opinion.

- John
_________________
I can confirm that I have received between 0 and 499 National Security Letters.
Back to top
View user's profile Send private message
saellaven
l33t
l33t


Joined: 23 Jul 2006
Posts: 646

PostPosted: Thu Oct 31, 2013 1:58 pm    Post subject: Reply with quote

I use LVM for all of my non-root directory needs but specifically keep root (and /boot for that matter) as traditional partitions specifically to avoid an initramfs because it's just another complication and potential point of failure.

I've been configuring my own kernels since the mid-90s, tuning them to my systems as needed. I don't need a one size fits all solution that throws everything into a module "just in case" and as we are not a binary distro, where the kernel, initramfs and core boot packages like lvm, get upgraded in sync with each other, having things fall out of sync due to an upgrade not warning about the need to rebuild initramfs becomes more likely and may cause boot failure.
_________________
Ryzen 3700X, Asus Prime X570-Pro, 64 GB DDR4 3200, GeForce GTX 1660 Super
openrc-0.17, ~vanilla-sources, ~nvidia-drivers, ~gcc
Back to top
View user's profile Send private message
ulenrich
Veteran
Veteran


Joined: 10 Oct 2010
Posts: 1480

PostPosted: Thu Oct 31, 2013 2:58 pm    Post subject: Reply with quote

I voted "No - other"

I wouldn't trust genkernel crafted. I trust upstream:
dracut -H initrd.new <kernel-version>
which dynamically slims down to what is needed and takes just a few seconds.

I just do that with my kernel-make script for validation purpose.
After a new kernel is perfectly configured I mostly don't need it any more and I rename the symlink:
mv kernel.new kernel.tested ; rm initrd.new
to happily boot my "tested" grub menu without any initrd.
Back to top
View user's profile Send private message
666threesixes666
Veteran
Veteran


Joined: 31 May 2011
Posts: 1248
Location: 42.68n 85.41w

PostPosted: Thu Oct 31, 2013 3:13 pm    Post subject: Reply with quote

i dont use them, though am not opposed to them either....

pro, compile modules rather than IN for required early boot system stuff. speeding up the kernel by having it act as an index of drivers rather than a hoarders mess.

con, takes longer to load an initrd then hand off to real kernel.

con, must rebuild initrd with new kernels.

con, more entries to boot loaders, more places to fail.

i did like uvesafb for my nvidia setup and i did use initramfs compiled into the kernel at that time. i havent been able to get it working for a while now though. when it did work it was amazing.
Back to top
View user's profile Send private message
lost+found
Guru
Guru


Joined: 15 Nov 2004
Posts: 509
Location: North~Sea~Coa~s~~t~~~

PostPosted: Thu Oct 31, 2013 3:26 pm    Post subject: Reply with quote

Yes (Other - please explain):

On 2 desktops: for fbsplash only. It can be done without initramfs, but then the boot splash will start a few seconds later.
On 1 headless server: no initramfs ("No, I do my own kernel so it's just more work").

All Gentoo, of course.

Cheers!


Last edited by lost+found on Thu Oct 31, 2013 9:41 pm; edited 1 time in total
Back to top
View user's profile Send private message
Randy Andy
Veteran
Veteran


Joined: 19 Jun 2007
Posts: 1148
Location: /dev/koelsch

PostPosted: Thu Oct 31, 2013 7:31 pm    Post subject: No (Other - please explain) Reply with quote

I also voted for, No (Other - please explain) cause:

I don't have a setup on any of my gentoo boxes, which require the need of a initramfs.
I prefer customized kernels for my specific needs and hardware cause:

- It boot faster, ok, only a little :wink:
- Its smaller, contains only the needed modules and as follow of this, compiles much faster without building hundreds of not needed modules every time, and I Use regular nearly every RC.
- It works more stable cause none conflicting automatisms as follow of failing auto-recognition of modules, so no blacklisting needed as a workaround for this (kms conflicts with some graphics drivers came to mind, lots of alsa trouble with hardware and hdmi codec conflicts and as longer I'm thinking about, as longer the list would get).
- It works more more secure. Remembering some vulnerable in the past of typical generic kernels I'm not affected by, cause I didn't had the reason for that configured cause I didn't need it.
- I don't like or need splash screens, it hides helpful information if something goes wrong.
- Bootloader menue entries are shorter and simpler.

- thanks to root=PARTUUID= for the kernel command line since kernel >= 2.6.37, compiled with CONFIG_EFI_PARTITION=y support and GPT partitioned storage devices, I don't need initramfs or grub2 to identify external boot (like usb-) devices clearly.


I know, there exist some scenarios for the need of initramfs also on Gentoo systems. But mostly it will be used for generic systems without knowing the hardware it should be running on and that is why most distributions would use it.
But for me, as a gentoo user, customizing is my first approach, so why should I stop at the kernel, where everything begins :wink:

Best, Andy.
_________________
If you want to see a Distro done right, compile it yourself!
Back to top
View user's profile Send private message
mrsteven
Veteran
Veteran


Joined: 04 Jul 2003
Posts: 1938

PostPosted: Thu Oct 31, 2013 8:07 pm    Post subject: Reply with quote

I voted no (other) - actually I compile my own kernel, so it's more work AND something else that might go wrong. Currently I don't need one, but in principle I wouldn't mind trying one if I needed to. I read it's quite easy to build one with dracut, but I haven't tested it so far, so my opinion might change in the future.
_________________
Unix philosophy: "Do one thing and do it well."
systemd: "Do everything and do it wrong."
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 54028
Location: 56N 3W

PostPosted: Thu Oct 31, 2013 8:34 pm    Post subject: Reply with quote

I voted yes as I have root in LVM on raid5.

However, I found an interesting post the other day that suggested an early boot environment in /boot, so I may vote no in future.
It would be an initrd on a real partition, so its not loaded by the boot loader
_________________
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
lost+found
Guru
Guru


Joined: 15 Nov 2004
Posts: 509
Location: North~Sea~Coa~s~~t~~~

PostPosted: Thu Oct 31, 2013 9:37 pm    Post subject: Re: No (Other - please explain) Reply with quote

Randy Andy wrote:
...
- I don't like or need splash screens, it hides helpful information if something goes wrong. ...


Yes, that's the down side. But there's a verbose mode too, which still shows console output with the eye candy as background. F2 switches between silent and verbose mode.
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Thu Oct 31, 2013 10:03 pm    Post subject: Reply with quote

I voted no (other) - as I don't need it.
If I needed it I would use one.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
mrbassie
l33t
l33t


Joined: 31 May 2013
Posts: 770
Location: over here

PostPosted: Thu Oct 31, 2013 11:50 pm    Post subject: Reply with quote

No. I only recently realised I didn't need one. Lots still to learn.
Back to top
View user's profile Send private message
Aiken
Apprentice
Apprentice


Joined: 22 Jan 2003
Posts: 239
Location: Toowoomba/Australia

PostPosted: Fri Nov 01, 2013 1:06 am    Post subject: Reply with quote

I went with do my own kernel and the ramfs is more work. Any driver needed to boot the various computers I have compiled into the kernel and like not having any slash screen. That covers normal use.

At the moment there is only 1 case where I use an initrd. If needed can use pxe to boot and the initrd contains enough to be used as a rescue image or to bootstrap a new image on a computer. That initrd is a normal fs made from either debootstrap or ROOT=image emerge .......
_________________
Beware the grue.
Back to top
View user's profile Send private message
Randy Andy
Veteran
Veteran


Joined: 19 Jun 2007
Posts: 1148
Location: /dev/koelsch

PostPosted: Fri Nov 01, 2013 10:57 am    Post subject: Re: No (Other - please explain) Reply with quote

lost+found wrote:
Randy Andy wrote:
...
- I don't like or need splash screens, it hides helpful information if something goes wrong. ...


Yes, that's the down side. But there's a verbose mode too, which still shows console output with the eye candy as background. F2 switches between silent and verbose mode.


Sure, but how do you know when to switch to the verbose mode, to see a specific error you wouldn't notice after its passed through and also not on your running OS.
Example: RTC faults, which prevent updating the hardware clock came to mind. After the network connection has been established and the ntp request has been done your OS clock shows the right time, but if you uses NVRAM/ACPI wakeup events, they won't be updated.

So using a splash screen, is more a question of taste regarding the design of your OS, no matter if you can switch it on or off for the worst case. :wink:
_________________
If you want to see a Distro done right, compile it yourself!
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sat Nov 02, 2013 2:12 am    Post subject: Reply with quote

I'm about to vote No as I've been using these (minor) patches to initscripts, to run upstream udev without an initramfs (and split /usr) for the last 2 years.

But I think this is a silly poll, given that Gentoo explicitly states not using an initramfs is unsupported, and we've had lots of propaganda in the last 2 years to explain why we simply must, and that our systems are not running correctly if we don't. The latter is complete BS, ofc, but given that it's unsupported, and explicitly so, I really would expect a massive percentage to be using one. Thus the poll is simply confirming that people are running Gentoo, imo.

OFC that won't stop it being used to prove that the initramfs adoption has been successful, and that Gentoo are right not to worry about it, since only a tiny percentage of "haters" are not using one.

edit: kk seen the percentages: give it some time, and I bet you the figures will skew the other way. Or this will be a "non-representative sample" since it was mostly "haters" who posted, ie those who have self-selected not to use an initramfs are more likely to vote in a poll like this. Unless ofc the figures do skew towards that, in which case it will be seen as even more conclusive proof that this is a fuss about nothing ("even with the selection bias, there are still 70% of the forums saying they use an initramfs.")

IOW people will always believe what they want to believe. Just like someone will no doubt state I am doing (it's not based on experience, no..)
Back to top
View user's profile Send private message
The Doctor
Moderator
Moderator


Joined: 27 Jul 2010
Posts: 2678

PostPosted: Sat Nov 02, 2013 3:34 am    Post subject: Reply with quote

"No, I do my own kernel so it's just more work"

This basically sums it up for me. Since it is easy enough to build a kernel that is able to boot on its own, it is simply superfluous to add an initramfs.

As far as I'm concerned its a tool. I would use one (and in fact have done so in the past) to boot special setups, such as an encrypted root. Always use the right tool for the right job.
_________________
First things first, but not necessarily in that order.

Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box.
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sat Nov 02, 2013 4:56 am    Post subject: Reply with quote

The Doctor wrote:
As far as I'm concerned its a tool. I would use one (and in fact have done so in the past) to boot special setups, such as an encrypted root. Always use the right tool for the right job.

Agreed: if you're doing anything unusual with rootfs you need one, same as you needed an initrd before.

Or if you have some need for udev to localmount (the case we always discuss is a bluetooth keyboard; that's the only one I've ever seen mentioned, ever since one developer mentioned his device needed it to boot, or rather if he wanted to interact with grub it did; but the point stands: you may have some device which is critical.) So no-one's ever said they're no use: just that they are only required in specific circumstances, or if you are on a bindist, which has a kitchen-sink kernel that loads whichever modules are needed in the initramfs, or you are yourself trying to build a generic "boot-anywhere" kernel (such as for a live-disk.)

Split /usr is not one of those cases: the only requirement is that local drives (the equivalent of localmount) are mounted before udev starts; and that's still a requirement in an initramfs. You just have the relevant drivers in there too, plus whatever else you need to get that far like lvm, mdadm etc. And of course you need to keep those in sync with the main system (so it's not just a case of only building it when you upgrade kernel.)

All perfectly scriptable of course.
Back to top
View user's profile Send private message
Havin_it
Veteran
Veteran


Joined: 17 Jul 2005
Posts: 1240
Location: Edinburgh, UK

PostPosted: Sat Nov 02, 2013 9:13 am    Post subject: Reply with quote

steveL wrote:
I'm about to vote No as I've been using these (minor) patches to initscripts, to run upstream udev without an initramfs (and split /usr) for the last 2 years.

But I think this is a silly poll, given that Gentoo explicitly states not using an initramfs is unsupported, and we've had lots of propaganda in the last 2 years to explain why we simply must, and that our systems are not running correctly if we don't. The latter is complete BS, ofc, but given that it's unsupported, and explicitly so, I really would expect a massive percentage to be using one. Thus the poll is simply confirming that people are running Gentoo, imo.

OFC that won't stop it being used to prove that the initramfs adoption has been successful, and that Gentoo are right not to worry about it, since only a tiny percentage of "haters" are not using one.

edit: kk seen the percentages: give it some time, and I bet you the figures will skew the other way. Or this will be a "non-representative sample" since it was mostly "haters" who posted, ie those who have self-selected not to use an initramfs are more likely to vote in a poll like this. Unless ofc the figures do skew towards that, in which case it will be seen as even more conclusive proof that this is a fuss about nothing ("even with the selection bias, there are still 70% of the forums saying they use an initramfs.")

IOW people will always believe what they want to believe. Just like someone will no doubt state I am doing (it's not based on experience, no..)


Well, FWIW I promise I'm not in the pay of the shadowy Gentoo overlords in their ivory tower ... but then, I would say that, wouldn't I :wink:

Seriously though, could you point me to some of this "propaganda"? If that is how devs/infra feel, it's news to me, but my community engagement level is perhaps not all it could be.

You state that your setup is one that would ordinarily require an initramfs per current baselayout/udev, and clearly you disagree with infra on that point, but really that's more to do with upstream than Gentoo, isn't it? And Gentoo does have at least one alternative (gudev?) so they are hardly just quietly toeing the line set by Poettering. I wonder if you are letting your own position colour your view a bit; because honestly I think the results are quite balanced so far, so I don't feel it was a "silly" poll at all.
Back to top
View user's profile Send private message
asturm
Developer
Developer


Joined: 05 Apr 2007
Posts: 8931

PostPosted: Sat Nov 02, 2013 10:13 am    Post subject: Reply with quote

I did not perceive such propaganda as well. Never have I used an initramfs in all those years.
Back to top
View user's profile Send private message
TomWij
Retired Dev
Retired Dev


Joined: 04 Jul 2012
Posts: 1553

PostPosted: Sat Nov 02, 2013 11:01 am    Post subject: Reply with quote

No (Other - please explain): I care about boot time because I need to reboot / kexec often; for me to use one, would be if there is a good reason to.
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sat Nov 02, 2013 11:03 am    Post subject: the ML Reply with quote

Havin_it wrote:
Well, FWIW I promise I'm not in the pay of the shadowy Gentoo overlords in their ivory tower ... but then, I would say that, wouldn't I :wink:

Hehe.
Quote:
Seriously though, could you point me to some of this "propaganda"? If that is how devs/infra feel, it's news to me, but my community engagement level is perhaps not all it could be.

I'd rather not; I realise that sounds weak, so let me apologise upfront: I should not have called your poll silly, and I was unwise to get into this.
The propaganda I'm referring to is the constant repetition of upstream systemd pages, as if they were somehow the gospel truth: if you want examples please PM me, and I'll search some out, if you cannot find them; I don't want to point to specific instances, as that will derail your poll into YAF argument about systemd and arm-twisting and so on. My bad. Suffice it to say that both the "gospels according to Lennart" have been fully debunked, but no we keep hearing about them, and people think "it's on freedesktop.org it must be true" whereas fd.o as a project repo is nothing more than a clique of developers who guard access.

This went a bit long; sorry about that: I figured what you're really after is insight into how to interact with the dev list, should you decide to become more involved. So take this with a pinch of salt, as it's just my experience: I can't speak for anyone else, so I don't.
Quote:
You state that your setup is one that would ordinarily require an initramfs per current baselayout/udev, and clearly you disagree with infra on that point, but really that's more to do with upstream than Gentoo, isn't it? And Gentoo does have at least one alternative (gudev?) so they are hardly just quietly toeing the line set by Poettering. I wonder if you are letting your own position colour your view a bit.

No I'm not, but that's because my position is in fact based on a poll I did a few years ago about a (global) server USE flag, and not at all on the udev situation.

I got all kinds of justification why the results were invalid, many of them basically disparaging users, afair, and the others presenting pseudo-statistical arguments. When I made the point that if an overwhelming majority of 200+? (can't mem) people vote one way, it's a clear indication of some sort of consensus (after debunking the "well, no one else would vote on it" argument) they all shut up and just let the thread die.
Nothing changed of course, even though it was only ever meant to be a "maintainer-decides" if they want to use it or not. Instead the "policy" continued, so someone who wanted eg: mysql client, would have to compile the whole server, despite the build-system explicitly supporting client-only builds, and even if the maintainer thought that would be a good idea. So you get client or server USE flags, and no simple one-shot consistency for server admins vs desktop clients. They're all local (which is fine: you can have a local.desc for a global flag, but they pretty much always mean the same, just the opposites according to package; except when they don't eg gdb.)

No-one could explain the rationale in a manner that made much sense; ultimately afaic developers are here to serve users, not the other way round. Or we get some new ones, or the distro dies eventually as it implodes into a clique of "users are the problem" idiots, imo. It goes in phases: after a period of egregiously bad behaviour, everyone makes a fuss of being user-friendly, and pretends no-one ever insulted users, and let's forget about the users we banned, while pretending our devs were spotless.

You may well hear that developers are corrected "off-list" or "privately". The same does not apply to users, by any means. Unfortunately that merely leads to a permanent record of users being insulted by devs with impunity, and no-one saying a word, whereas users are rounded-upon by many devs. Which as you can imagine, is a very bad impression to give, especially to anyone considering more involvement, and perusing the list to check out what they might be getting into. It has got better, though. They just do the same thing in other media, and still feel free to insult users, but will brook nothing that could be considered an insult against a developer, even if it's not about the developer but a criticism of his work. (and it usually is a he: there's very few female devs; christel and musikc are
the only ones I know of, and they retired ages ago.)

Either they think that's how to behave, usually accompanied by "well I'll just have to win their respect, it's an initiation rite" or they turn away and seek a less poisonous atmosphere to work in. Neither are very conducive to a healthy collaborative atmosphere. I've seen, and interacted with another FLOSS project that does it very differently over the last few years: KDE, where developers act the same as users on these forums; they want everyone to do well, and are not into ego-trips, just great software. AFAICT they appear to be a fair bit older than Gentoo devs, but that's just an impression, from a small sample-size. I've got much more experience of interacting with Gentoo devs, though. I wish I didn't, right now as they can, like anyone can, be incredibly nasty. They do this much more with "formal processes" which ofc they are in a better position to know, influence, and indeed are treated specially as discussed.

Don't get me wrong: it's in the nature of FLOSS that arguments are carried out in the open, exactly the same as the forums. However ime moderators on the forums don't really care what your name or title is: they'll call it like they see it, irrespective of status, and are quite happy to discipline any user without fear nor favour. And since that always starts with a warning in the thread reported, it's always done openly. Well, it used to be; haven't seen that happen for a while now.

To my mind, that's why Gentoo has lasted as long as it has: the forums are the heart of the community, along with IRC, where you also get pretty good moderation, in the larger channels. It's impossible to guarantee across all of them, since some are small, so if you don't get on with the main person or two, you're screwed and better find somewhere else, and some simply aren't official gentoo channels, so you have no recourse in any event. But out of the two only the forums are permanent.

It's also why the developers, after a massive 6 month effort to get community consensus, pulled out of having proctors at the last minute with one flip "I don't think we need all those rules." Simply put the most experienced mods were forum mods, and there were quite a few on that team. A few years later they merged devrel and userrel, and that's comrel now; not sure but I don't think there's many forum mods on it; one ex-forum mod who is now a developer that I know of. So effectively the community is ultimately policed not by the team with the most experience of that very task, but those who are politically acceptable to other devs, since that's how comrel are appointed: just like a herd.

Justice has to be seen to be done, to be effective: sheltering your devs from the consequences of their own actions merely inculcates a sense of invulnerability, and "abuse of power comes as no surprise." It's human nature, which is why developers are supposed to be held to a higher standard. If you read back over the list, I'd be very surprised if you find many examples of devs doing that. I certainly have not in the last what 7 years, nor in the list I read before that (from the beginning.) I only did that, as so many users complained about what a den of vipers it was, and I just couldn't believe that. I saw what they meant when I read the list, and considered it as a whole, not just when someone was feeling like they needed to prove their "good-guy" credentials.

I haven't read it much for the last month: but like I said it goes in phases, and you need to consider it over a longer period. Not that I haven't said things I regret, nor posted too much at the beginning: if you're not used to it, when someone responds to you, you feel you should respond back (after all it's only polite, right?) Sadly many of the responses are simply petty point-scoring, and if you respond in anything approaching the same tone, you will be slapped about by as many devs as can hit Send before someone magnanimously says "OK we've pummelled him enough."

So, if you're going to post, restrict it to two a day. It will save you a lot of hassle, and no-one will be able to accuse you of very much, presuming you are not obnoxious. If they're getting to you, realise that, and back away from the keyboard: come back in a day or two, and more often than not you'll find someone else has posted in support of your position, albeit in different words. And if not, you can make your point a lot more coolly, and typically there's more you can pull into your post; meaning you can respond to several mails at once, which makes your post more interesting, since you're actually pulling strands together.

Getting back to the technical point (both aspects of software development are critical) no there is no technical reason why we cannot start udev after localmount, considering the nature of the distro. We are lucky: we can still use the same criteria as before: if you have root on lvm/raid/encrypted/NFS, or you have a device which udev must initiate before you can localmount, you need an initrd (or rather an initramfs nowadays.) If not, you can quite happily start udev after localmount, assuming you have your mobo, rootfs and HDD modules built-in to kernel. device-mapper is also required for lvm, though i cannot advise on exactly what you need built-in for your setup: that's what lspci -k, the web, and #gentoo are for ;)

Basically if you do the install, configure and build a kernel, and get it to reboot to a login, with all your partitions mounted (which is the first step every time I install gentoo) then you're set. Typically people have their NIC builtin as well, and don't have any other modules for mobo or hard-disk controller. Why would they?

So yes, this is strictly a distro decision, as the patches show: there is no patching at all of udev itself, only two minor patches to initscripts. That could be supported just as easily, and yet you have people like WilliamH consistently presenting 2 or 3 other options to the Council as an exhaustive list, as if there is no other option. It's simply untrue, and he knows it: since he's well aware of the patches, and so is the rest of the dev community. They basically told me to STFU about it, when all I tried to do was get both "sides" to collaborate instead of argue, since the "problems" of early-init are exactly the same whether you use an initramfs or not. You need everything that you're going to use to mount other partitions: either in initramfs, or on rootfs.

They didn't like hearing that, for some reason. *shrug* I don't post to the list very much now at all, and that kind of behaviour is the reason why. Very occasionally I post when I get fed up with the extraordinary level of BASH incompetence, but they stopped sending post-commit related messages. Ho hum. That was about the only time it was particularly interesting as you saw discussion about actual product. You get it occasionally when they post eclasses for review; which they only do because they have to. Drobbins knew exactly what he was doing when he setup the rules for that dev ML. It must have torn him apart to come back to Gentoo, only to find it had descended into what he was fleeing when he set it up. He left again within 2 months istr.

Note: mysql is just an example, don't infer anything at all about anyone's position from it. Also as ever the above is just my opinion: make your own mind up, and remember: Gentoo is a very loose collective of individuals. About the only time they display unaniminity, IME, is when they're beating up on a user for "insulting" a developer. So tread carefully (they're not very good at detaching their ego from their work: most of them don't even know what that means), and take refuge in the forums, when you're feeling got at.
And when you're not: spread the love :-)

Oh one last thing: the dev ML is not the developers' list: it's the community's list to discuss development, which anyone can post to (stay on-topic though: it's not the same as the forums, where mods will split off threads as they see fit: you cannot do that on an email list.) The devs have a private "gentoo-core" list; but must use the public list to discuss GLEPs, eclasses and any cross-distro changes they want to make. Also bear in mind, that many, but by no means all, developers are young (the influx tends to be from students at University.)


Last edited by steveL on Sat Nov 02, 2013 11:42 am; edited 8 times in total
Back to top
View user's profile Send private message
TomWij
Retired Dev
Retired Dev


Joined: 04 Jul 2012
Posts: 1553

PostPosted: Sat Nov 02, 2013 11:23 am    Post subject: Re: Advice to users before you post to dev ML. Reply with quote

steveL wrote:
Ultimately afaic developers are here to serve users, not the other way round. Or we get some new ones, or the distro dies eventually as it implodes into a clique of "users are the problem" idiots, imo.


Developers are users, they come from out of the community; if the community (including developers) has no further interest, then either neither or both sides are to be blamed for that but not a single side. There is no such thing as "the users are the problem" because then a developer would simply be referring to oneself as well; the same goes for recruiting new people, it's not only the users that don't step up but also the developers making it not interesting enough to join for them. And at moments like that when a distro runs empty; lowering the recruitment barrier would perhaps help, but that goes at the cost of some quality which could get some other users and develpers angry. And then history just repeats until the distro and its community (including developers) either dies due to ego or gives up caring about unnecessary details and gets out of its regression...


Last edited by TomWij on Sat Nov 02, 2013 11:38 am; edited 3 times in total
Back to top
View user's profile Send private message
GFCCAE6xF
Apprentice
Apprentice


Joined: 06 Aug 2012
Posts: 295

PostPosted: Sat Nov 02, 2013 11:27 am    Post subject: Reply with quote

ulenrich wrote:
--snip--
[..]dracut[..]
--snip--


Big +1 for dracut. I made a bit of a fool of myself a while back parroting some negative crap I read to a friend of mine who is an Arch user, needless to say he ended up giving me a little beat down and had me give dracut a test. I have to say I was pleasantly surprised how quick and easy it is. Just one of those things where when you don't use something (the two boxes I look after which do need one do it all automatically) you become rather ignorant to facts or reality, unfortunately.
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 54028
Location: 56N 3W

PostPosted: Sat Nov 02, 2013 11:56 am    Post subject: Reply with quote

steveL,

+1 mostly

steveL wrote:
... Gentoo is a very loose collective of individuals ...

I would say (and have said before) "... Alpha Males ..."
For the avoidance of doubt, I'm probably the oldest active Gentoo dev.

Also, Gentoo is not the workplace. Its not possible to force volunteers to do things, you have to invoke their self interest.
*rel can do some wrist slapping but there is really only one sanction they have to apply force and thats to fire the dev in question.
There is nothing in between.
_________________
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
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page 1, 2, 3, 4  Next
Page 1 of 4

 
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