How about hosting tutorial on http://gentoo-wiki.comali3nx wrote:Just a brief update on the tuitorial files. I'll host them on my own webserver for the meantime. If anyone has a copy of the make.conf from my old file stash on ali3nx.info email me a copy. I've allready hosted one but the use flags have varied to some degree and i would like to maintain some margin of consistancy with the use flags. My email address is available in my profile. Happy new year to all and thanks
Installing Gentoo - The Developers Method - Stage1 and NPTL
Message
*drumroll*kowal wrote:How about hosting tutorial on http://gentoo-wiki.comali3nx wrote:Just a brief update on the tuitorial files. I'll host them on my own webserver for the meantime. If anyone has a copy of the make.conf from my old file stash on ali3nx.info email me a copy. I've allready hosted one but the use flags have varied to some degree and i would like to maintain some margin of consistancy with the use flags. My email address is available in my profile. Happy new year to all and thanks
http://gentoo-wiki.com/Installing_Gentoo_-_The_Gentoo_Developers_Method_with_NPTL_and_2.6_from_Stage1
*/drumroll*
Much thanks to bsodmike for getting it started.Compiling Gentoo since version 1.4
Thousands of Gentoo Installs Completed
Emerged on every continent but Antarctica
Compile long and Prosper!
Thousands of Gentoo Installs Completed
Emerged on every continent but Antarctica
Compile long and Prosper!
I've been tinkering with Rac's robust installation method. I'm a little confused by a couple of things in your post.hielvc wrote:... So instead of doing a bootstrap you...Code: Select all
emerge --oneshot glibc binutils gcc > for a regular, non ~x86 <Do this for ~x86 if you want nptl> emerge -C linux-headers && emerge --oneshot linux26-headers && emerge --oneshot glibc binutils gcc <Then both groups do this> emerge system -e
So the new super duper one liner would look like this.Code: Select all
env-update && source /etc/profile && emerge --oneshot --nodeps gcc-config && emerge linux26-headers && emerge --oneshot glibc binutils gcc && emerge -e system && emerge syslog-ng xinetd grub hotplug coldplug vixie-cron reiserfsprogs reiser4progs sysfsutils udev dhcpcd && emerge --nodeps acpid ntp && rc-update add syslog-ng default && rc-update add net.eth0 default && rc-update add vixie-cron default && rc-update add xinetd default && rc-update add sshd default && rc-update add hotplug default && rc-update add coldplug default && rc-update add acpid default
first, the first code listing includes, but the second code listing (one-liner) doesn't include the "emerge -C linux-headers" step. IME with a 2004.3 Stage 3 tarball, you have to emerge -C the linux headers, as failing to do so results in a failed emerge of linux26-headers with linux-headers acting as a blocker. (this step isn't necessary if you start with a Stage 1 tarball). hasn't anyone else noticed this, or is nobody else using a Stage 3 tarball?
second, I see in the first code snippet above that you've replaced the bootstrapping script with equivalent toolchain building steps. These make sense. In the second code listing, the super-duper one-liner adds "emerge --oneshot --nodeps gcc-config" before the toolchain is built. i'm not sure why this step exists in the second code listing but not in the first, and i'm hoping someone can explain why its being used in the second code listing. thx.
Yes, I'm sure some people out there have noticed, but they're all lurkers who don't comment at all on the difference.Bob P wrote:the second code listing (one-liner) doesn't include the "emerge -C linux-headers" step. IME with a 2004.3 Stage 3 tarball, you have to emerge -C the linux headers. hasn't anyone else noticed this, or is nobody else using a Stage 3 tarball?
It's to get around a long-standing bug in gcc way back from october - the sanity check errors: http://forums.gentoo.org/viewtopic.php? ... 06#1691306 But no one knows if it has been bugfixed yet.Bob P wrote:In the second code listing, the super-duper one-liner adds "emerge --oneshot --nodeps gcc-config" before the toolchain is built. i'm not sure why this step exists in the second code listing but not in the first, and i'm hoping someone can explain why its being used in the second code listing. thx.
An additional insertion i've only just thought of now, but is mentioned in rac's thread: emerge portage along with the toolchain in the first pass. I got sick and tired of seeing 4 NPTL installs restart the package count 23 packages into "emerge -e system", after portage was emerged. I can't tell if this will work, but I'm installing with this insertion into the one-liner now.
The new stage 3 reduced hassle no bootstrap super-duper one-liner
Code: Select all
env-update && source /etc/profile && emerge -C linux-headers && emerge linux26-headers && emerge --nodeps --oneshot gcc-config && emerge --oneshot glibc binutils gcc portage && emerge -e system && emerge syslog-ng xinetd grub hotplug coldplug vixie-cron reiserfsprogs reiser4progs sysfsutils udev dhcpcd && emerge --nodeps acpid ntp && rc-update add syslog-ng default && rc-update add net.eth0 default && rc-update add vixie-cron default && rc-update add xinetd default && rc-update add sshd default && rc-update add coldplug default && rc-update add hotplug default && rc-update acpid defaultthat's one of the differences between the Stage 1 and Stage 3 tarballs with this guide -- AFAIK the Stage 1 tarballs for 2004.3 come with linux26-headers, while the Stage 3 tarball doesn't.kimchi_sg wrote:Yes, I'm sure some people out there have noticed, but they're all lurkers who don't comment at all on the difference.Bob P wrote:the second code listing (one-liner) doesn't include the "emerge -C linux-headers" step. IME with a 2004.3 Stage 3 tarball, you have to emerge -C the linux headers. hasn't anyone else noticed this, or is nobody else using a Stage 3 tarball?
because ali3nx's guide is using the Stage 1 tarball, we've reached a point of divergence when we start talking about Rac's methods.
after emerging gcc aren't we supposed to do another mandatory "etc-update && source /etc/profile"? i notice that this has never appeared after emerging gcc in the one-liner...kimchi_sg wrote:It's to get around a long-standing bug in gcc way back from october - the sanity check errors: http://forums.gentoo.org/viewtopic.php? ... 06#1691306 But no one knows if it has been bugfixed yet.
...
The new stage 3 reduced hassle no bootstrap super-duper one-linerCode: Select all
env-update && source /etc/profile && emerge -C linux-headers && emerge linux26-headers && emerge --nodeps --oneshot gcc-config && emerge --oneshot glibc binutils gcc portage && emerge -e system && emerge syslog-ng xinetd grub hotplug coldplug vixie-cron reiserfsprogs reiser4progs sysfsutils udev dhcpcd && emerge --nodeps acpid ntp && rc-update add syslog-ng default && rc-update add net.eth0 default && rc-update add vixie-cron default && rc-update add xinetd default && rc-update add sshd default && rc-update add coldplug default && rc-update add hotplug default && rc-update acpid default
No, the stage 1 tarballs do not come with any form of kernel header files at all. Just extract a stage 1 to a temp directory, chroot into it and env-update, then do an emerge info and you will see what i mean. The headers will be shown as not installed, I verified it myself.Bob P wrote:AFAIK the Stage 1 tarballs for 2004.3 come with linux26-headers, while the Stage 3 tarball doesn't.
while i've been building my base layout i've been i've been sifting through the volume of posts on this thread, and i've found your 3-liner. it looks like a very good approach, and its very similar to the approach i have been taking. thanks for pointing it out.Hobbit_HK wrote:I remember posting a 3-liner so you could actually use the new CFLAGS gcc3.4 has before rebuilding your system.. It's here somewhere, I think after the first supr-duper-oneliner
There were a couple of reasons why I pointed out rac's post. For those of you who are newer rac is one very bright bulb. I saw his post and I didnt want it to just disapear. The point of his post concerns this thread. This Howto is is one of the better ones around for ending up with a bleeding edge system. Before I found this thread I had stoped doing stage1s and with a homebrew of experince, magic, three or four other threads and a stage3 I was doing what ali3n so straight forwardly put togather.
The point of this Howto is to get your system built and going the fastest way possible. Most of us have seen a whole bunch of problems recently starting from a stage1 and bootstrap. Some of them could have been avoided by starting with a stage3 and according to rac even when we dont have probs there are
This can be avoided and there are usually less problems by " emerge system -e" from a statge3. So why not go this route?
As to kimchi_sg's new super dupper one liner looks good.
EDIT puttin words that escaped
The point of this Howto is to get your system built and going the fastest way possible. Most of us have seen a whole bunch of problems recently starting from a stage1 and bootstrap. Some of them could have been avoided by starting with a stage3 and according to rac even when we dont have probs there are
There are some 80+ packages in a stage1ball that are not listed in /var/db/pkg. Why? When you do your "emerge system", you would want your new toolchain to be used to compile all software. If portage sees that a particular version from the stageball is still current, it will omit it. The solution that somebody apparently chose was to make portage forget that most of this software is installed at all, which has the unfortunate side effect of making portage be unable to clean it when your "emerge system" finishes.
This can be avoided and there are usually less problems by " emerge system -e" from a statge3. So why not go this route?
As to kimchi_sg's new super dupper one liner looks good.
EDIT puttin words that escaped
Last edited by hielvc on Tue Jan 04, 2005 7:23 pm, edited 1 time in total.
Or from this tucows miror,i found it with google yesterday.Oktane wrote:Namesys still down, but if somebody needs that Reiser4 patch against 2.6.10, you can get it from here.
blank screen
hi all, thanks for the great tutorial. I follow all steps in this tutorial and evrything went well, but after I reboot my system start and did all the loading well ,when it comes to the login I got a blank screen with only the cursor.
I did emerge ati-drivers ,emerge x11-xorg. still the same problem .
Any idea guys? please help
I did emerge ati-drivers ,emerge x11-xorg. still the same problem .
Any idea guys? please help
unfortunately, the process of performing a Stage 1 Install on a Stage 3 Tarball is more complicated that it appears at first glance. I mentioned this in rac's thread that gave birth to this idea.hielvc wrote:This can be avoided and there are usually less problems by " emerge system -e" from a statge3. So why not go this route?
As to kimchi_sg's new super dupper one liner looks good.
the one flaw that i see in every take on the whole Stage 1 on 3 approach that we've seen so far is that they're all performed from a Stage 1 mindset. the contents of a Stage 3 tarball are different. in addition to changing the way you bootstrap the system, you also have to consider where you're going with the newly emerged version of GCC.
Hobbit-HK made a tremendously important point that nobody seems to have picked up on -- that you have to break the one-liner apart when you install a new version of GCC. He made some good points about taking pause to think about the CFLAGS statements in /etc/make.conf before getting carried away with the one liner. there are definitely some worthwhile benefits to be gained in that respect.
one other thing that nobody else seems to have noticed is that all of the one liners i've seen to date fail to actually install the GCC 3.4.3 compiler after emerging it! if you follow the one-liner as it is written, the entire "emerge -e system" et.seq. is executed using GCC 3.3.4, even though we're all assuming that its being done with 3.4.3. if you have any doubts about this, do a real Stage 3 install on a blank HD and see what i mean -- if you look at the output of "gcc-config -l" after emerging 3.4.3 you'll see something that looks like this:
Code: Select all
#gcc-config -l
[1] i386-pc-linux-gnu-3.3.4 *
[2] i386-pc-linux-gnu-3.4.3
[3] i386-pc-linux-gnu-3.4.3-hardened
[4] i386-pc-linux-gnu-3.4.3-hardenednopie
[5] i386-pc-linux-gnu-3.4.3-hardenednosspCode: Select all
gcc-config 2into the one-liner to change the default compiler before getting carried away with "emerge -e system". its probably a good idea to recompile all of the compiler tooklit components before going there too. what do you think?
Bob P: Thanks for alerting us to that!
But i've just d/led the stage 1 for bootstrapping my pentium 3 laptop. And i dun want to re-d/l stage 3 again.
Btw, thanks for mentioning me in the credits in your new 'stage 1 from stage 3' tutorial.
Do u know if the extra 'gcc-config 2' step is required for a stage 1 tarball bootstrap as well?
But i've just d/led the stage 1 for bootstrapping my pentium 3 laptop. And i dun want to re-d/l stage 3 again.
Btw, thanks for mentioning me in the credits in your new 'stage 1 from stage 3' tutorial.
Do u know if the extra 'gcc-config 2' step is required for a stage 1 tarball bootstrap as well?
you know, i'd been reading through rac's thread it seemed that in that thread and in this one, we were all tweaking ali3nx's methods to adapt them to the Stage 3 install from the Stage 1 mindset. that is to say, i don't think that you have to do the "gcc-config 2" step in a Stage 1 install because (has Hobbit pointed out) you emerge GCC onto a blank system. at least that's how i remember it. because we're all used to doing Stage 1 installs, we've been biased by the lack of necessity for this step. as a result, the need for replacing GCC 3.3.4 with 3.4.3 when doing the 1 on 3 wasn't intuitively obvious to us. it had totally escaped me until i actually decided to quit reading about "virtual installs" and perform a real install on a real PC. in getting ready to do that, i re-read all 21 pages of this post from beginning to end, and then it hit me -- alot of things have to be different to get the Stage 1 on 3 install right.kimchi_sg wrote:Bob P: Thanks for alerting us to that!
But i've just d/led the stage 1 for bootstrapping my pentium 3 laptop. And i dun want to re-d/l stage 3 again.
Btw, thanks for mentioning me in the credits in your new 'stage 1 from stage 3' tutorial.
Do u know if the extra 'gcc-config 2' step is required for a stage 1 tarball bootstrap as well?
the differences became significant enough that i kept track of them on a notepad as i went through everything over the New Year's holiday. because there were enough differences, it seemed only logical to put the revised method in its own thread. to me at least, it seemed better to keep this thread about doing a real Stage 1, and to create the other thread for doing a Stage 1 on 3. hopefully, both processes will be less confusing if we don't have crossover in the instructions.
in regard to mentioning you guys in the tut -- most of the answers to how to do a Stage 1 on 3 the right way were already posted here, so i thought it was only fair to give credit to the guys who came up with the ideas.
Btw, i don't know if anyone has actually tried a Stage 1 on 3 following my directions yet, but so far the results have been great -- but because i did it on a Pentium 133 (yikes!) its still doing the emerge -e world and it will probably be stuck there for a couple of more days...
i need to hit a computer show later this month and pick up a couple of more boxes for testbeds.I use diferent tipe of instalation of nptl with gcc 3.4
emerge libperl
Code: Select all
...
..
mirrorselect -a -s4 -o |grep 'GENTOO_MIRRORS=' >> /mnt/gentoo/etc/make.conf
cp -L /etc/resolv.conf /mnt/gentoo/etc/resolv.conf
mount -t proc none /mnt/gentoo/proc
chroot /mnt/gentoo /bin/bash
env-update
source /etc/profile
emerge --sync
nano -w /etc/make.conf
##Edit after change to new profile 2005.0
rm -f /etc/make.profile
ln -sf /usr/portage/profiles/default-linux/x86/2005.0 /etc/make.profile
mkdir /etc/portage
nano -w /etc/portage/package.keywords
echo "sys-devel/gcc ~x86" >> /etc/portage/package.keywords
echo "sys-libs/glibc ~x86" >> /etc/portage/package.keywords
echo "sys-libs/libstdc++-v3 ~x86" >> /etc/portage/package.keywords
echo "sys-kernel/linux26-headers ~x86" >> /etc/portage/package.keywords
emerge -C linux-headers
emerge --oneshot --nodeps linux26-headers
## and bootstrap
cd /usr/portage
scripts/bootstrap.sh
## check gcc
gcc-config -l
## and emerge system
emerge system
ls /usr/share/zoneinfo
ln -sf /usr/share/zoneinfo/GMT /etc/localtime
etc ....
Last edited by mudrii on Thu Jan 06, 2005 1:40 am, edited 2 times in total.
ithreads is a pretty helpful USE flag too. IIRC, emerging libperl with the nptl flag set, but no ithreads flag set, results in a warning message that specifically asks for the ithreads USE flag to be set, and then recompiling libperl. of course, you won't see this message scroll away if "emerge libperl" is buried in a one-liner.
