| View previous topic :: View next topic |
| Author |
Message |
arkaine
n00b
Joined: 23 Jun 2004
Posts: 62
Location: Ohio
|
 Posted: Thu Nov 11, 2004 11:50 pm Post subject: |
 |
|
dumb post, deleted
Last edited by arkaine on Fri Nov 12, 2004 7:48 am; edited 1 time in total |
|
| Back to top |
|
 |
slycordinator
Advocate
Joined: 31 Jan 2004
Posts: 3065
Location: Korea
|
| Posted: Fri Nov 12, 2004 7:23 am Post subject: |
|
|
| arkaine wrote: | | I'll try that jmusits, though I do believe I had already done that...Also is it just me or does xorg HATE USE="nptl" ? I've used the NPTL guide by snooper (I believe that's his name) and this one and both times it just seems stuck on the unpacking the 5 or 6 source files. |
nptl should not affect anything related to unpacking files; nptl is only related to thread support. So how did you conclude that the problem was the USE flag? |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Fri Nov 12, 2004 11:01 pm Post subject: Re: Installing Gentoo - The Developers Method - Stage1 and N |
|
|
| ali3nx wrote: | To create file systems:
| Code: | mkreiserfs /dev/hda1 --format 3.6
mkreiserfs /dev/hda3 --format 3.6
mkswap /dev/hda2
swapon /dev/hda2 |
|
ali3nx, i want to ask you about using reiser fs on /dev/hda1. i've done this Stage 1 install following your directions on two machines. both of them eventually croaked when the /boot partition achieved 100% utilization of space, reportedly because of the journal file.
i created the /boot partition on both of my machines using Reiser FS 3.6 as instructed here. i made both partitions +32M in size as recommended in the Gentoo Install Handbook.
in short order, both of these machines were no longer able to do much of anything useful because all of the free space on the /boot partition became used up (invisibly, i might add) by the reiser journals.
so i have to admit, i am a bit lost in regard to why this problem is coming up. several people have advised me on other threads that Reiser FS should not be used on small partitions because the journaling system will eat up the partition's free space. they have recommended ext2fs for the /boot drive instead.
so i'm wondering how you're able to keep your systems running with Reiser FS 3.6 on the boot partitions. are you doing something different, perhaps making the size of the partitions bigger? or am i missing something?
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
Deranger
Veteran
Joined: 26 Aug 2004
Posts: 1215
|
| Posted: Sat Nov 13, 2004 1:13 am Post subject: |
|
|
32 MB for a /boot is way too small, especially for journalling filesystem. You should create over 100 MB /boot just like ali3nx said in the tutorial, I have myself 128 MB /boot, using ext-3.
| ali3nx wrote: |
I highly recommend using +100M or larger boot partitions for a 2.6 kernel install. 2.6 kernels use allot more space than 2.4 and hey.. why redo it later.
|
|
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
|
| Back to top |
|
|
monkey89
Guru
Joined: 08 Mar 2004
Posts: 596
|
| Posted: Sat Nov 13, 2004 2:34 am Post subject: |
|
|
| Oktane wrote: | 32 MB for a /boot is way too small, especially for journalling filesystem. You should create over 100 MB /boot just like ali3nx said in the tutorial, I have myself 128 MB /boot, using ext-3.
| ali3nx wrote: |
I highly recommend using +100M or larger boot partitions for a 2.6 kernel install. 2.6 kernels use allot more space than 2.4 and hey.. why redo it later.
|
|
You could turn on every feature and probably still be under 10MB from what I see. My kernels are 4 MB each with enough features to run my system well, and I can't imagine you have so much hardware that you'd manage to make a kernel bigger than 15 MB at most.
A journaled filesystem is overkill for a /boot partition. The handbook recommends ext2 for a reason - it doesn't take up nearly as much space on the drive. 32 MB is and always has been fine for me, and you should have ample room for multiple kernels should the need come up. |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Sat Nov 13, 2004 2:55 am Post subject: |
|
|
the deceptive part of running reiserfs is that you don't see the size of its journals growing to consume space on the drive. this happens transparently to the user.
in my case, 10% of my boot partition was used by kernels and such, and my queries into how much space on the drive my kernels and such were occupying resulted in very low numbers, along the lines of 10% of the disk space. meanwhile reiser was invisibly attempting to use 100% of the drive for its journals. because 10% + 100% = 110%, it creates a problem.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
ephilli2
n00b
Joined: 18 Sep 2004
Posts: 27
|
| Posted: Sat Nov 13, 2004 6:58 am Post subject: Reiser4? |
|
|
| I'm about to start a new install on a Dell 600m. I want to use Reiser4; how do the instructions change? Is there a certain live cd I should use? lxnay? The guide doesn't seem as scary as the first time I read it, good job. |
|
| Back to top |
|
|
slycordinator
Advocate
Joined: 31 Jan 2004
Posts: 3065
Location: Korea
|
| Posted: Sat Nov 13, 2004 7:35 am Post subject: Re: Reiser4? |
|
|
| ephilli2 wrote: | | I'm about to start a new install on a Dell 600m. I want to use Reiser4; how do the instructions change? Is there a certain live cd I should use? lxnay? The guide doesn't seem as scary as the first time I read it, good job. |
The Ixnay livecd works for reiser4; I installed with it using this guide along with another similar guide.
If the ONLY thing you're changing is using reiser4 here's what you'd have to modify:
1) When creating the new file system, simply issue the command "mkfs.reiser4"; mkreiserfs is for the older versions not reiser4.
2) Obviously emerge a kernel that supports reiser4 and include reiser4 support. |
|
| Back to top |
|
|
UncleAelfrich
n00b
Joined: 10 Nov 2004
Posts: 6
Location: Los Angeles, CA
|
| Posted: Sat Nov 13, 2004 8:56 am Post subject: |
|
|
Ali, I used your method here to install a system on my Athlon 1300 machine. It worked well. Thanks a lot. I had no real problems UNTIL I had to configure it for static IPs. It might be helpful if you added a note in your tutorial about the different steps one would take in configuring for static IP. Again, thanks for your help.
John Chapman
Los Angeles, CA |
|
| Back to top |
|
|
Deranger
Veteran
Joined: 26 Aug 2004
Posts: 1215
|
| Posted: Sat Nov 13, 2004 11:10 am Post subject: |
|
|
| monkey89 wrote: | | Oktane wrote: | 32 MB for a /boot is way too small, especially for journalling filesystem. You should create over 100 MB /boot just like ali3nx said in the tutorial, I have myself 128 MB /boot, using ext-3.
| ali3nx wrote: |
I highly recommend using +100M or larger boot partitions for a 2.6 kernel install. 2.6 kernels use allot more space than 2.4 and hey.. why redo it later.
|
|
You could turn on every feature and probably still be under 10MB from what I see. My kernels are 4 MB each with enough features to run my system well, and I can't imagine you have so much hardware that you'd manage to make a kernel bigger than 15 MB at most.
A journaled filesystem is overkill for a /boot partition. The handbook recommends ext2 for a reason - it doesn't take up nearly as much space on the drive. 32 MB is and always has been fine for me, and you should have ample room for multiple kernels should the need come up. |
That's true, my kernels are under 1.5 MB  32 MB is enough for non-journalling filesystems. Handbook also recommends ext-3 for root, for a reason?  It's the crappiest filesystem (well, ext-2 is even more worse) for a root. |
|
| Back to top |
|
|
ali3nx
l33t
Joined: 21 Sep 2003
Posts: 722
Location: Winnipeg, Canada
|
| Posted: Sat Nov 13, 2004 12:44 pm Post subject: |
|
|
my reasoning for recommending 100MB boot partitins are
1) remote servers are something you do not want to re-engineer
2) your kernels are very valuble. making room for grub + 5+ kernel
updates gives you both redundancy from the journal and security
by lowering the frequency required to delete stale kernel images that
occupy space. If you do have a problem with a corrupted boot
filesystem space then what are you to do? loose your kernels?
3) reduces nessesity required of administration to tend to your filesystems
freeing time for other tasks. Filesystem maintainance can be very time
consuming across an enterprise.
Good planning will allways reward you in the long run. Despite the extra 10 megs of your priceless 200gb hard disk the journal will take to add redundancy to your systems. It is up to you.. but who will be watching the game on saturday with no redundacy planning
-:edit:- my kernels are rarely ever more than 1.6 megs. but when a boot filesystem restoration is required from a tarball due to the unlikely chance that journalled reiserfs ever does become corrupt the extra 50 megs that isn't in use can become an asset in a hurry. Plan for crunch time and you wont find yourself in the board room with donald pointing the finger at you
_________________
Compiling Gentoo since version 1.4
Thousands of Gentoo Installs Completed
Emerged on every continent but Antarctica
Compile long and Prosper! |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Sat Nov 13, 2004 7:55 pm Post subject: |
|
|
ali3nx, thanks for that well thought out post that clearly shows how amateurs like me can learn alot of useful techniques from IT professionals like you.
i just happen to have fallen into a n00bie knowledge gap where i didn't know enough to plan ahead for the need to: 1) keep extra kernels and to provide the space to accomodate them, 2) account for space taken up by the Reiser journals, and 3) compile or modularize ext2fs support into the kernels in case i needed to fall back on it because of partition size restrictions.
in my case, i'm spending saturday working on the PC instead of watching the game. 
thanks again for starting this great instructional thread.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Sat Nov 13, 2004 8:18 pm Post subject: |
|
|
| UncleAelfrich wrote: | | Ali, I used your method here to install a system on my Athlon 1300 machine. It worked well. |
John, when you did your bootstrapping, did you encounter the problem where perl did not compile as a dependency of emerge system? (see page 7 of this thread).
that was a definite show-stopper for me. i logged a bug report into bugzilla about it, but it doesn't seem to have been addressed. i'm wondering if the issue has been resolved even though there has been no update to bugzilla.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
UncleAelfrich
n00b
Joined: 10 Nov 2004
Posts: 6
Location: Los Angeles, CA
|
| Posted: Sun Nov 14, 2004 9:04 pm Post subject: |
|
|
nope ... had no problems with that one ... although I did decide to recompile and configure my mouse into the kernel
I would appreciate any help on configuring the boot process to load my static IPs and the gateway information |
|
| Back to top |
|
|
ali3nx
l33t
Joined: 21 Sep 2003
Posts: 722
Location: Winnipeg, Canada
|
| Posted: Sun Nov 14, 2004 9:47 pm Post subject: |
|
|
since baselayout seems to have defaulted to a blank /etc/conf.d/net and a full example at /etc/conf.d/net.example you can merely copy the example to the usable file from within chroot.
| Code: | | cp /etc/conf.d/net.example /etc/conf.d/net |
_________________
Compiling Gentoo since version 1.4
Thousands of Gentoo Installs Completed
Emerged on every continent but Antarctica
Compile long and Prosper! |
|
| Back to top |
|
|
UncleAelfrich
n00b
Joined: 10 Nov 2004
Posts: 6
Location: Los Angeles, CA
|
| Posted: Mon Nov 15, 2004 12:41 am Post subject: |
|
|
Thanks, I had gotten that ... but when I put my gateway IP in
routes_eth0=( "default gw xxx.xxx.xxx.xxx" )
I get two red !! when the computer boots up. I assume that means that the boot process had some problem loading the .net file in terms of the gateway address. (The eth0 address works fine and I am able to ping the local network as well as the internet.) |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Mon Nov 15, 2004 9:11 pm Post subject: More dependencies on libperl... |
|
|
I'm in the process of doing another Stage 1/NPTL install, and I thought I'd provide feedback on how its going with the new 2004.3 Stage 1 tarball, as this may help ali3nx keep his HOW-TO up to date:
1. when execusing the "emerge -C linux-headers" command, i received an error message stating that no linux headers were present to be unmerged. maybe this step isn't necesary when using the new 2004.3 tarball.
2. with the new 2004.3 Stage 1 tarball, the bootstrapping procedure (scripts/bootstrap.sh) ran successfully without a hitch. (on a P3-800 it took 10 hours.) It seems that the problems that I had run into with the 2004.2 tarball (other apps being dependent on unmerged perl) have been appropriately addressed in the tarball.
3. there seem to be some unresolved portage issues. during the "emerge system" step, I am running into more dependency on perl issues. my emerge system failed after only a few of the 72 apps to be emerged had been completed. at this point i'm doing an "emerge libperl && emerge libperl" before I return to "emerge system". we'll see how that works out...
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
FastTurtle
Guru
Joined: 03 Sep 2002
Posts: 477
Location: Flakey Shake & Bake Caliornia, USA
|
| Posted: Tue Nov 16, 2004 2:32 am Post subject: |
|
|
I'm in the proccess of following a combination of this guide, the udev guide and doing a stage1 with the new 04.3 live cd minimal on a Celeron 2.4 laptop. Stage1 is already completed and things look to be faster already.
Thank god for the -fetchonly flag for emerge. I don't know what's been wrong with the internet the last week but things have been slow (10-30K/s) instead of my usual (70-300k/s). Hope to begin stage2 sometime tonight.
Note on use flags:
One thing I learned fairly early with 1.4_rc4 was to keep use flags to the absolute minimum during bootstrap. The only things I include are cpu specific in order to reduce dependency issues. Not fun to have X trying to build when bootstraping a system.
ali3nx:
Good points about planning for fs failure/corruption but as most laptops have pretty small drives, I'm sticking with ext2 for boot - 20 megs is plenty while the remaining partitions are ext3 with fairly large journals. Once I begin rebuilding the desktop, I may go ahead and use reiserfs on it for optimum speed.
Bob P:
It appears that the 04.3 minimal install cd has finally removed linux-threads. Check the size of the ISO, about 1 meg less then the 04.2 minimal.
Hope it means the devs are finally getting somewheres on cascading.profiles.
_________________
AsRock B550 Phantom Gaming 4
128GB 3200 Mhz memory
4x 4TB Sata - 2x 2TB Sata SSD - 4x 450GB SaS - 3x 900GB SaS - 72GB SaS for Boot
LSI 9211-8i in HBA mode for all of the SaS drives
Radeon 6800 (Non XT) for GPU |
|
| Back to top |
|
|
vangelis
n00b
Joined: 05 Jan 2004
Posts: 15
Location: St. Louis, USA
|
| Posted: Tue Nov 16, 2004 6:17 pm Post subject: going crazy... |
|
|
I am not sure I will be able to take it for too much more...
I am trying to put together a firewall/router/VPN box. I am using a 400MHz K6-2 system with 256MB RAM.
Anyway I started the 2004.3 stage1, followed the instructions precicely, (I had the same issue with the linux-headers not being there any more, but I just continued...) and after more than 24 hours... bootstrap finished.
Despite the fact that I kept seeing "nptl" and "posix" everywhere during compilation, when I entered /lib/libc.so.6 this is what I get (I have repeated the bootstrap process from scratch twice!!):
| Code: | GNU C Library 20041102 release version 2.3.4, by Roland McGrath et al.
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.4.3 (Gentoo Linux 3.4.3, ssp-3.4.3-0, pie-8.7.6.6).
Compiled on a Linux 2.6.8 system on 2004-11-15.
Available extensions:
GNU libio by Per Bothner
crypt add-on version 2.1 by Michael Glad and others
linuxthreads-0.10 by Xavier Leroy
The C stubs add-on version 2.1.2.
GNU Libidn by Simon Josefsson
BIND-8.2.3-T5B
libthread_db work sponsored by Alpha Processor Inc
NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
For bug reporting instructions, please see:
<http://www.gnu.org/software/libc/bugs.html> |
Here is my make.conf:
| Code: | CFLAGS="-O2 -march=k6-2 -fomit-frame-pointer -pipe -mmmx -m3dnow"
CHOST="i586-pc-linux-gnu"
CXXFLAGS="${CFLAGS}"
GENTOO_MIRRORS="http://gentoo.chem.wisc.edu/gentoo/ http://gentoo.mirrors.pair.com/ "
USE="nptl objc f77 fortran erandom pic"
ACCEPT_KEYWORDS="~x86"
|
The question now is: Do I have the posix threads ready to go, or something went wrong? I do not want to embark on the emerge system before I make sure that I have the posix enabled tools (god knows how many days that will take...) In any event I do have gcc-3.4.3 as the gcc-config -l shows:
| Code: | # gcc-config -l
[1] i386-pc-linux-gnu-3.3.4
[2] i586-pc-linux-gnu-3.4.3
[3] i586-pc-linux-gnu-3.4.3-hardened *
[4] i586-pc-linux-gnu-3.4.3-hardenednossp
|
By the way, while I was at it I switched to the hardened version since I might need it for the security features of the firewall.
Any ideas?
Thanks
vangelis |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Tue Nov 16, 2004 8:07 pm Post subject: Re: going crazy... |
|
|
| vangelis wrote: | | Anyway I started the 2004.3 stage1, followed the instructions precicely, (I had the same issue with the linux-headers not being there any more, but I just continued...) and after more than 24 hours... bootstrap finished. |
i do not think that "issues" with the execution of the command "emerge -C linux-headers" are anything to worry about. afaik that command is used to remove 2.4 headers from the bootstrap. if you read ali3nx's original post on this thread, the bootstrap scripts have changed quite a bit in the past month. i am of the impression that the 2004.3 installation procedure no longer requires that command, and that it can be deleted from ali3nx's instructions. this is probably just a case of the installation instructions in this thread not yet being fully updated to reflect all of the changes brought on by 2004.3. remember, its only been out for a few days now, and we've just pointed this "issue" out.
Ali3nx: for reference, here's the typical screen output following the "emerge -C linux-headers" command as-issued in your one-liner: | Code: | livecd portage # emerge -C linux-headers
--- Couldn't find linux-headers to unmerge.
>>> Unmerge: No packages selected for removal. |
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Tue Nov 16, 2004 8:22 pm Post subject: Re: going crazy... |
|
|
| vangelis wrote: | Despite the fact that I kept seeing "nptl" and "posix" everywhere during compilation, when I entered /lib/libc.so.6 this is what I get (I have repeated the bootstrap process from scratch twice!!):
| Code: | GNU C Library 20041102 release version 2.3.4, by Roland McGrath et al.
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.4.3 (Gentoo Linux 3.4.3, ssp-3.4.3-0, pie-8.7.6.6).
Compiled on a Linux 2.6.8 system on 2004-11-15.
Available extensions:
GNU libio by Per Bothner
crypt add-on version 2.1 by Michael Glad and others
linuxthreads-0.10 by Xavier Leroy
The C stubs add-on version 2.1.2.
GNU Libidn by Simon Josefsson
BIND-8.2.3-T5B
libthread_db work sponsored by Alpha Processor Inc
NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
For bug reporting instructions, please see:
<http://www.gnu.org/software/libc/bugs.html> |
|
FWIW, I see the exact same information when i issue that command on my PC. my question to you is, are you bothered by that? it says that linuxthreads-0.10 is installed, so i'm not sure that i understand your conderns.
| vangelis wrote: | | The question now is: Do I have the posix threads ready to go, or something went wrong? I do not want to embark on the emerge system before I make sure that I have the posix enabled tools (god knows how many days that will take...) |
Don't worry too much about how much time "emerge system" will take. The bootstrap is by far the most time consuming part of the install. For reference, I'm listing some of the elapsed times that were requried for some of the major operations during my last install on a P3-800. You shoud be able to extrapolate somewhat to your processor, based on MHz: | Code: | hr:min system commands
----- ------------------------------
10:15 scripts/bootstrap.sh
00:30 emerge system # 72 files to merge, crapped out with perl dependency at gettext-1.03
00:06 emerge libperl && emerge libperl
02:20 emerge system # 54 packages left |
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks
Last edited by Bob P on Wed Nov 17, 2004 8:12 am; edited 1 time in total |
|
| Back to top |
|
|
vangelis
n00b
Joined: 05 Jan 2004
Posts: 15
Location: St. Louis, USA
|
| Posted: Tue Nov 16, 2004 8:35 pm Post subject: |
|
|
I am sorry if I misled you Bob P. I also think that the missing linux-headers is not a significant problem. I just mentioned it as a side issue.
My main question is about the posix threads. Given the output I am getting, do I have them or not?I am not really bothered, but I am going through all this trouble to get them in my system (for all the advantages that they are claimed to have), so when I don't see it there I feel deceived...
Doing a little more research on the issue, I just realised that they can co-exist with the linuxthreads, if you don't specify "nptlonly" in the USE statement.
The new question is then, will the system always choose to use posix threads over the linuxthreads when possible?
If not, and you then force the posix threads by including nptlonly, will you break something in the system down the road?
Thank you all
vangelis |
|
| Back to top |
|
|
Burnce
n00b
Joined: 04 Oct 2004
Posts: 1
|
| Posted: Wed Nov 17, 2004 7:38 am Post subject: smp kernel not present on liveCD (2000.3 minimal x86) |
|
|
Hi,
I attempted to install gentoo 2004.3 witl gcc 3.4.3 and nptl today.
In the howto at the beginning of this forum it says to boot using the smp kernel. There is no such option on the x86 minimal liveCD.
I tried to proceed anyway, but bootstrap.sh failed. Do I need to install gcc3.3.x and then upgrade? I was trying to go directly for gcc 3.4.3. (maybe I was just naive but I edited /etc/portage/packages.keywords and tried to run bootstrap.sh from there).
Thanks for any help,
Burnce |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Wed Nov 17, 2004 8:09 am Post subject: |
|
|
you're right -- SMP is not a boot option on the 2004.3 live CD. you can just press the <CR> which is the equivalent of typing "gentoo" at the prompt.
if you follow the directions here (don't go editing anything that isn't in the instructions!) your bootstrap should go fine. but everyone does seem to be running into the gettext dependency on perl issue during emerge system. this may well be an error in the ebuild of gettext which fails to tell portage that it is dependent on perl.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
|
Documentation, Tips & Tricks
