| View previous topic :: View next topic |
| Author |
Message |
desultory
Administrator
Joined: 04 Nov 2005
Posts: 7448
|
 Posted: Wed Jan 21, 2009 9:46 am Post subject: |
 |
|
| slonocode wrote: | | This doesn't seem to address the issue that it seems a better solution in practice as well. Which creates the problem of why wouldn't gentoo official choose to use the better tool. Which is seemingly something to complain about. |
Your dislike of a tool does not mean it is not suited to its purpose nor does your liking another tool mean it is suited to the role it would need to fill. Before claiming the superiority of one tool over another take the time to check what each can do, at the moment only one of these has all of the necessary features. In short, either get the tool you like better to do what it needs to be able to to to fill the role you want it to fill or let others get on with what they are already doing. |
|
| Back to top |
|
 |
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Wed Jan 21, 2009 4:56 pm Post subject: |
|
|
| desultory wrote: | | As part of setting up the chroot: |
mount --bind /dev/ /mnt/gentoo/dev/ |
|
Two comments:
- That step is not shown in the instructions I was following: http://www.gentoo.org/doc/en/gentoo-x86-quickinstall.xml
- That step only fixes things while you're in the console. It doesn't fix the brokenness that occurs after a reboot(e.g. no initial console).
|
|
| Back to top |
|
|
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Wed Jan 21, 2009 6:34 pm Post subject: |
|
|
| grante wrote: | That step only fixes things while you're in the console. It
doesn't fix the brokenness that occurs after a reboot(e.g. no
initial console). |
That didn't come out quite right. What I meant was it only
fixed things while you're in the chroot'ed environment. It
doesn't fix /dev so that a reboot comes up correctly. |
|
| Back to top |
|
|
slonocode
Apprentice
Joined: 03 Jun 2002
Posts: 273
|
| Posted: Wed Jan 21, 2009 7:27 pm Post subject: |
|
|
| desultory wrote: | | slonocode wrote: | | This doesn't seem to address the issue that it seems a better solution in practice as well. Which creates the problem of why wouldn't gentoo official choose to use the better tool. Which is seemingly something to complain about. |
Your dislike of a tool does not mean it is not suited to its purpose nor does your liking another tool mean it is suited to the role it would need to fill. Before claiming the superiority of one tool over another take the time to check what each can do, at the moment only one of these has all of the necessary features. In short, either get the tool you like better to do what it needs to be able to to to fill the role you want it to fill or let others get on with what they are already doing. |
I didn't say I like or dislike either tool. But your posts don't address the fact that in practice one seems to put out pretty good stages every three days and one seems to have trouble doing it on a weekly basis. My hope was that you could address that perception or correct it if it's wrong. Talking in philosophical abstracts doesn't really do that. |
|
| Back to top |
|
|
desultory
Administrator
Joined: 04 Nov 2005
Posts: 7448
|
| Posted: Thu Jan 22, 2009 8:50 am Post subject: |
|
|
| grante wrote: | That didn't come out quite right. What I meant was it only
fixed things while you're in the chroot'ed environment. It
doesn't fix /dev so that a reboot comes up correctly. |
Having just checked stage3-x86-20090121.tar.bz2, I can confirm that it does in fact have a properly populated /dev/, technically ./dev/, and that MAKEDEV is present. In short, it has no obvious reason to fail as you described. Lacking further information, diagnosis of the problem you have described can proceed no further.
| slonocode wrote: | | I didn't say I like or dislike either tool. |
You claimed that one "seems a better solution in practice as well" while giving no technical grounds for that assessment, indicating that you prefer it to the other. As such, relative like and dislike seems an appropriate description of your stated opinion.
| slonocode wrote: | | But your posts don't address the fact that in practice one seems to put out pretty good stages every three days and one seems to have trouble doing it on a weekly basis. |
On the off chance you meant that seriously, according to their respective operators both are used to periodically generate stages from the same repository using the same package manager and tool chain to do it, as such their products should logically be rather similar. As for schedules, even weekly releases for stages is past the point of diminishing returns, more often only makes sense if there are major changes being made or there is active testing being done as part of the release system, given his most recent comments in the bug tracker that seems to not be a practical concern of Daniel Robbins.
| slonocode wrote: | | My hope was that you could address that perception or correct it if it's wrong. |
I represent neither those releasing either set of images nor those developing either system to produce them, so to address your concerns I will simply recommend that you actually read the introductory documentation for both systems, note that according to its own documentation the one you describe as superior simply lacks functionality which is required by Gentoo. To summarize, better or not it is not a viable solution at this time.
| slonocode wrote: | | Talking in philosophical abstracts doesn't really do that. |
Claiming that one thing is superior to another while strongly implying a total lack of experience with either generally indicates trolling. |
|
| Back to top |
|
|
slonocode
Apprentice
Joined: 03 Jun 2002
Posts: 273
|
| Posted: Thu Jan 22, 2009 10:04 am Post subject: |
|
|
You have provided links to pages using untrusted security certificates. Outstanding.
I describe it as superior because 1 person can generate stages regularly and often and the other is more typical of the gentoo norm in releasing sporadically and much less often. As someone who seemed to know more about it I was hoping you could speak to that observation. Clearly you are more interested in a semantics battle. |
|
| Back to top |
|
|
slackline
Veteran
Joined: 01 Apr 2005
Posts: 1300
Location: /uk/sheffield
|
| Posted: Thu Jan 22, 2009 11:16 am Post subject: |
|
|
| slonocode wrote: | You have provided links to pages using untrusted security certificates. Outstanding.
|
Have you looked at where the links point to? Its Gentoo's Bugzilla, if you aren't already registered and accepted the certificates then you should do as a lot of the activity in Gentoo development hinges on this.
And on a personal note (i.e. its my opinion)...
Why not stop bitching and picking holes in the developers work and instead get yourself registered at Bugzilla and start help squashing some bugs, its a far more productive contribution. 
_________________
"Ubuntu" - an African word meaning "Gentoo is too hard for me". |
|
| Back to top |
|
|
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Thu Jan 22, 2009 3:24 pm Post subject: |
|
|
| desultory wrote: | | grante wrote: | That didn't come out quite right. What I meant was it only
fixed things while you're in the chroot'ed environment. It
doesn't fix /dev so that a reboot comes up correctly.
|
Having just checked stage3-x86-20090121.tar.bz2, I can confirm
that it does in fact have a properly populated /dev/,
technically ./dev/, and that MAKEDEV is
present. In short, it has no obvious reason to fail as you
described.
|
So others have actually installed using that stage3 and not had
problems? I've done two installs with it, and each time I ended
up having to fix /dev/ after the first reboot
| Quote: | Lacking further information, diagnosis of the problem you have
described can proceed no further. |
I think the main issue that I've seen is the lack of
/dev/console on startup. That results in a message from the
kernel something like "unable to open initial console" and then
you get no messages from anything while the various init
scripts are running. Then /etc/issue gets re-written with info
on how to do a "mount -bind" and a "cp -L" to fix the /dev
entries.
I don't understand why other people wouldn't have seen this
problem.
--
Grant |
|
| Back to top |
|
|
desultory
Administrator
Joined: 04 Nov 2005
Posts: 7448
|
| Posted: Fri Jan 23, 2009 1:59 am Post subject: |
|
|
| slonocode wrote: | | I describe it as superior because 1 person can generate stages regularly and often and the other is more typical of the gentoo norm in releasing sporadically and much less often. As someone who seemed to know more about it I was hoping you could speak to that observation. |
Aside from speculation by unassociated parties, I have encountered no indication that there are any technical difficulties restricting release frequency by either effort.
| slonocode wrote: | | Clearly you are more interested in a semantics battle. |
Hardly. My interest here is to reduce the amount of disinformation being spread and to provide assistance to those who have encountered actual problems.
| grante wrote: | I think the main issue that I've seen is the lack of
/dev/console on startup. That results in a message from the
kernel something like "unable to open initial console" and then
you get no messages from anything while the various init
scripts are running. Then /etc/issue gets re-written with info
on how to do a "mount -bind" and a "cp -L" to fix the /dev
entries. |
The first things to check are that you have adequate permissions to unpack the tarball properly, that is it is done as root, and that the partition is mounted so as to allow device nodes to be created on it when the tarball is unpacked, you can check with mount if it lists nodev among the options used for that partition you have found the problem. |
|
| Back to top |
|
|
genstorm
Advocate
Joined: 05 Apr 2007
Posts: 2238
Location: Austria
|
| Posted: Fri Jan 23, 2009 2:16 am Post subject: |
|
|
| grante wrote: | So others have actually installed using that stage3 and not had
problems? |
Being one of those, I cannot report any errors.
_________________
backend.cpp:92:2: warning: #warning TODO - this error message is about as useful as a cooling unit in the arctic |
|
| Back to top |
|
|
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Fri Jan 23, 2009 3:37 am Post subject: |
|
|
| genstorm wrote: |
| grante wrote: | | So others have actually installed using that stage3 and not had problems? |
Being one of those, I cannot report any errors. |
So where did your /dev/console file come from?
It's not in the stage3 tarball, and at no point in the installation
instructions is one created.
udev will create one, but before that it's going to be missing. Without a
/dev/console file I get an error message about how somebody (init?) is
"unable to open initial console",and then all ofthe startup logging is missing.
Are you booting up with an initial ramdisk as root? |
|
| Back to top |
|
|
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Fri Jan 23, 2009 3:45 am Post subject: |
|
|
| desultory wrote: |
The first things to check are that you have adequate permissions to unpack the tarball properly, that is it is done as root, and that the partition is mounted so as to allow device nodes to be created on it when the tarball is unpacked, |
There _is_ no /dev/console node in the tarball. That's the problem.
If there's no /dev/console in the stage3 tarball, where in the install process does
it get created?
Are you claiming there is a /dev/console entry in your copy of the tarball?
Here's what's in mine:
| Code: |
$ tar tvf stage3-i686-20090114.tar.bz2 | fgrep ./dev
drwxr-xr-x root/root 0 2009-01-13 21:20 ./dev/
-rw-r--r-- root/root 0 2009-01-13 21:20 ./dev/null
lrwxrwxrwx root/root 0 2009-01-13 19:29 ./dev/MAKEDEV -> ../sbin/MAKEDEV
-rw-r--r-- root/root 0 2009-01-13 19:29 ./dev/.keep
drwxr-xr-x root/root 0 2009-01-13 19:29 ./dev/shm/
-rw-r--r-- root/root 0 2009-01-13 19:29 ./dev/shm/.keep
drwxr-xr-x root/root 0 2009-01-13 19:29 ./dev/pts/
-rw-r--r-- root/root 0 2009-01-13 19:29 ./dev/pts/.keep
$
|
And yes, I've verified the file hash values. |
|
| Back to top |
|
|
grante
n00b
Joined: 21 Jan 2009
Posts: 7
|
| Posted: Fri Jan 23, 2009 4:39 am Post subject: |
|
|
| grante wrote: | | I've tried a couple installs using stage3-i686-20090114.tar.bz2. |
| grante wrote: | After one upacks stage3-i686-20090114.tar.bz2 in /mnt/gentoo,
how does one get the /mnt/gentoo/dev directory
populated? |
| desultory wrote: | Having just checked stage3-x86-20090121.tar.bz2,
|
Nice bit of redirection, that. You slipped it right by me the
first time.
You know, a simple "yes, the /dev entries in 10090114 were
indeed missing, but the problem has already been fixed and
20090121 is OK" would have been sufficient.
There was no need to try to make me look stupid by knocking
down and bayoneting a straw man.
| desultory wrote: |
I can confirm that it does in fact have a properly populated
/dev/, technically ./dev/, and that MAKEDEV is present.
|
No, technically, /dev/MAKEDEV is _not_ sufficient. Without a
/dev/console, things just don't work right. Disagree with that
all you want, but it's still true. IIRC, /dev/null is also
required.
Oh, and there's really no need for 5000+ /dev entries in the
stage3 tarball when udev is being used. It's just a waste of
inodes since they all get hidden as soon as udev starts. I
suppose it does give you a better chance of booting into a
working system if udev somehow fails completely.
--
Grant |
|
| Back to top |
|
|
desultory
Administrator
Joined: 04 Nov 2005
Posts: 7448
|
| Posted: Fri Jan 23, 2009 9:36 am Post subject: |
|
|
| grante wrote: | Nice bit of redirection, that. You slipped it right by me the
first time.
You know, a simple "yes, the /dev entries in 10090114 were
indeed missing, but the problem has already been fixed and
20090121 is OK" would have been sufficient.
There was no need to try to make me look stupid by knocking
down and bayoneting a straw man. |
Believe it or not, that was entirely oversight on my part, I just checked the most recent and listed the one I checked. Having just checked the same tarball you specified, stage3-i686-20090114.tar.bz2, I can confirm that it is lacking necessary device nodes, though I do not know why off hand.
| grante wrote: | No, technically, /dev/MAKEDEV is _not_ sufficient. Without a
/dev/console, things just don't work right. Disagree with that
all you want, but it's still true. IIRC, /dev/null is also
required. |
Given the invitation, I hereby disagree with that all that I want. The purpose of the MAKEDEV script is to create the necessary device nodes, invoking it properly would have solved the problem.
| grante wrote: | Oh, and there's really no need for 5000+ /dev entries in the
stage3 tarball when udev is being used. It's just a waste of
inodes since they all get hidden as soon as udev starts. I
suppose it does give you a better chance of booting into a
working system if udev somehow fails completely. |
Agreed. |
|
| Back to top |
|
|
Flarkis
Apprentice
Joined: 06 Aug 2008
Posts: 197
Location: Tallinn, Estonia
|
| Posted: Sat Jan 24, 2009 8:22 pm Post subject: |
|
|
just did an instal with the new builds...they work great...thanks team
_________________
Veni, Vidi, Vici
[I came, I saw, I conquered] |
|
| Back to top |
|
|
Kasumi_Ninja
Veteran
Joined: 18 Feb 2006
Posts: 1825
Location: The Netherlands
|
| Posted: Sun Feb 01, 2009 9:20 pm Post subject: |
|
|
This looks great. Maybe this is a good time install Gentoo again  It would be a good idea to announce the availability of new stages each week.
_________________
Please add [solved] to the initial post's subject line if you feel your problem is resolved. Help answer the unanswered |
|
| Back to top |
|
|
-Craig-
Guru
Joined: 03 Jun 2004
Posts: 333
|
| Posted: Fri Mar 06, 2009 7:02 pm Post subject: |
|
|
This is so great! I really like the autobuilds, the "old" method really sucked after a few months, because you had to update >100 packages and met some blockers on the way...
Thanks! |
|
| Back to top |
|
|
iamben
Apprentice
Joined: 10 May 2004
Posts: 232
|
| Posted: Tue Mar 17, 2009 6:04 pm Post subject: |
|
|
| edit:nm! |
|
| Back to top |
|
|
bob doe
Apprentice
Joined: 11 Jun 2007
Posts: 207
|
| Posted: Wed Mar 25, 2009 6:36 pm Post subject: |
|
|
| nm, found them. |
|
| Back to top |
|
|
1shot1kill
n00b
Joined: 09 Feb 2006
Posts: 16
Location: Akron, Ohio
|
| Posted: Tue Jul 21, 2009 10:35 pm Post subject: |
|
|
Weekly Builds are great!
Anyway we can get the git port built into the weekly builds? |
|
| Back to top |
|
|
nightmorph
Developer
Joined: 23 Jan 2005
Posts: 1385
Location: SoCal
|
| Posted: Wed Jul 22, 2009 4:35 am Post subject: |
|
|
You want the tarballs to include dev-util/git? No chance. The stage tarballs contain only what you need to install a bare-minimum command-line system along with a toolchain. We don't put extra packages in there, only the essentials. This is so that your installed environment is as flexible as possible, so that you can customize it to what you want, rather than some arbitrary pre-defined set of desktop or server packages, for example.
We give you the bare essentials for downloading and compiling your preferred Linux environment, no more, no less.
_________________
<UzzaDead> What is CONFIG_USB_MON?
<petteyg> A Jamaican USB configuration?
dirtyepic: "We have more cupholders."
GDP || PR |
|
| Back to top |
|
|
|
Gentoo Chat
