Gentoo Networking Problem New to Gentoo....need help desperately

[qwijibow 0]

Gentoo wont damage your hardware, its just that compiling from scratch will run your CPU at 100% for a few hours, if your cooling isnt up-to it, it may overheat.

you missed "-fomit-frame-pointer" from the CFLAGS, but this should just reduce performance a little, and should not cause any instability.

the porgram to monitor cpu tempreture is lm_sensors, you need to have compiled support for your monitoring chip on the kernel.

when you check the md5sums, test the cd's md5 againsed the .iso.md5 on the server.
the corruption may have occured dureing download.

memtest86.com http://www.memtest86.com/

this is a confusing error !

Faulty hardware would normally cause a kernel panick, and memory dump.
So i would guess its not your hardware.

And Bugs in the compiler / KDE should cause a seg-fault error...
and your not getting this.

is it possable you made a mistake somwhere ?

Gentoo is probably the second hardest distro to install, right behind Linux From Scratch.

Maybe try a stage3 install ?
i stage3, all important and critical software like the compiler, and glibc are pre-compiled.
yo just need to compile Xorg and KDE.

Otherwise... Keep hacking through, get dirty, find the problem and fix it..( the fun bit, lol )

Or... go back to Slackware.

I Think gentoo is fantastic,

But if you dont enjoy doing some dirty work, and tinkering with the inner workings of linux, It can become quite annoying at times like these.

EDIT: oohhhh, and about your nvidia problem...
assuming you have compiled and loaded the driver (load with modprobe nvidia)

and you get the sae error starting X, have a look in /dev/ for the files /dev/nvidia0 and /dev/nvidiactl.

if they dont exist, run the following as root.

mknod /dev/nvidiactl c 195 255 mknod /dev/nvidia0 c 195 0chown root.video /dev/nvidia*chmod 0660 /dev/nvidia*

then restart X

/etc/init.d/xdm restart

[abhiram 0]

I'm almost at the point of giving up now . I ran memtest and it gave me errors in excess of 1000! Does this mean my memory is completely screwed? I haven't had any memory problems before neither with windows nor slackware. Of course there is that occasional reboot in windows (maybe once a week or once a fortnight) but that's attributed more to windows than the hardware right?About the md5sum ... that's exactly what i did. I wrote the image on the CD, then re-imaged it and then cross-checked it's md5sum with the server's. Now, it's consistently giving me an error whenever it tries to emerge kdelibs. I tried it once again just now and it gave me a segmentation-fault. I do enjoy tinkering around with it ... but there's no point if I keep running into a wall. So far, I've been at it in excess of 4 days and it still hasn't worked nor is it showing any promise of working. Maybe there's some base problem in the stage1 and stage2 installs that I did. Would it be better if I started over and did a stage3 install?

[abhiram 0]

Well... that's that for gentoo. I tried a stage 3 install today and still errors. Well, maybe there is a small problem in my hardware but I don't think I can do anything about it. Anyways, everything else works fine and I did learn a lot of things from the gentoo install . I think I'm an expert at compiling kernels now . Thanks for your help qwijibow, but I'm going back to slackware. Long live binary installs!

[qwijibow 0]

I ran memtest and it gave me errors in excess of 1000!

Your hardware is faulty. No doubt about it.

It seems your hardware is working fine under low stress.
But memtest86 is designed to run your ram at full speed, doing read/write operations at every clock cycle.

Its most likely your Ram, or possably (but un-likely) the CPU.

Are you over-clocking your hardware ???

Did you tweak your ram timings ?
i used to get crashes, and a bad memtest86 because of a bios bug that ran my ram with bad timing settings..

i have Kingston ValueRAM.. my motherboard defalts the ram timings to 1& 2-2-6-3.
i fixed the stability issues by chaging ram timing to the Kingston approved 2T 3-3-8-3.

maybe there is a small problem in my hardware but I don't think I can do anything about it

You may want to run memtest86 a few more times.
If its always the same area of ram that fails (the same address range)
you can config the Linux kernel to not use that area of ram.

ANYWAYS......

You are right to give up gentoo.
Compiling the kernel, and things like glibc under faulty hardware will give you a vey very un-stable system.

At least your ram is running fine under low-stress.

Do you ever run memory intensive programs under windows / slackware ?
for example memory benchmarks, or high performance games ?

[abhiram 0]

I don't recall overclocking my hardware ... there isn't any option for overclocking given in the manual. One thing I forgot to mention, I have two 256MB RAM cards with different bus speeds. When I bought the 2nd one I didn't think about matching the speeds, I just bought it because it was at an all time low. Can we change the timings or something?My system would restart if I played Doom 3 or HL 2 but, that's only on rare occasions ... it wouldn't restart much if I stopped all the services and resources. Quake 3 runs fine without any glitches. I did try benchmarking my graphics card with 3dmark, but that's different I guess .Anyways, I've spent way too much time on this ... so back to slackware 10.2 tonight .

[qwijibow 0]

I have two 256MB RAM cards with different bus speeds

possably the source of the porblem...

Lets say for example you have 1 stick of PC2700 and one of PC3200, and you run your buss at PC3200 (400 mhz FSB) then the slower stick is overclocked, and possably causing the trouble.

When you have 2 different ram sticks with different frequency's you have to set yor board to run your ram at the slower sticks speed, for the above example, 333Mhz FSB (PC 2700)

If you have your Mobo set the the fester RAM clock speed, set it the speed of the slower ram stick... You may find that this fixes your crash problems with HL2.

[abhiram 0]

I guess you're right. I downloaded cpu-z and then checked the hardware specifications.

 

Apparently, I've got one PC2100 and one PC2700. For some very strange reason, there does not appear to be any provision, either on the motherboard (jumpers) nor in the BIOS, to change the FSB or clock speed. I've even searched online and downloaded the lates BIOS update from ASUS.

 

Here are some screenshots of my hardware:

 

 

 

 

Now, I've got some friends over here ... and the FSB of my M/B is apparently 400. So I could find someone and exchange RAMs with him. Which one would you say I should change? the 166MHz one or 133MHz?

[qwijibow 0]

okay.. im not an expert.. but here's what it looks like to me....You motherboard has a Max FSB of 400 Mhz, but is running at 333mhz.333Mhz FSB will run your ram at 166mhzYou have 2 sticks of ram, one is 133mhz, one is 166mhzSOOOOOit looks like the motherboard's bios has detected that it needs to slow your FSB for your ram.here is a crazy theory ( that may just prove correct)Maybe your BIOS only probed your first RAM stick for its clock speed, and then assumed that the other ram stick would be the same.An then clocked the FSB to match the ram in slot 1.Your fastest ram stick is in slot 1, with your slowest in slot 2.Anyways... What if... Open up your case, ans swap over the ram sticks.So your 133Mhz stick is in slot1 and your 166Mhz is in slot 2.Maybe your motherboard will then set your FSB to 266mhz (which is the correct speed for the slower stick)If this works, it will slightly under-clock your faster ram stick, but hopfully wou should get your stability back.Maybe half-life 2 will stop crashing.Alternativly....If you can get your hands on some PC3200 Ram, then your motherboard will be able to run at its maximum 400mhz FSB. (assuming your CPU can handle 400mhz FSB)

[abhiram 0]

Well, I don't know how to say this but ... I guess I've been wrong the whole time. I was under the impression that if slack and winxp were working, there shouldn't be a significant problem with the hardware. But now, I just checked each RAM individually using memtest and it seems that one of my RAMs is screwed up. I'll have to get it replaced but I don't think that will happen for another 2 months. Thanks for your help ... I really appreciate you taking your time out for replying so regularly. I really want to try gentoo out and I'll give one last attempt tonight with my 256MB RAM (I checked it thoroughly and there are no problems). I'll post back on the results. Thanks once again for your time .

[qwijibow 0]

No problem, answering forum questions is a welcome break from job seeking [ just left university ] :sIt seems that one of the ram sticks was over-clocked, but surviving.Maybe a gentoo stage 1 full compile for several hours on an over clocked ram slot must have killed it.Good luck and next time, make sure your ram chips frequency's match

[abhiram 0]

Well, I'm done .

I must say, it was quite easy after I removed the faulty RAM. A straight, simple, by-the-book, easy install. Since I read the handbook 3 times already, it wasn't a big deal. Started the installation at 9PM with a stage 3 and by 1:00 AM, I had X running, and alsa configured and was listening to songs using mpg123 while I was finishing the base configuration. I gave the command 'emerge kdebase-startkde kicker' and went to sleep and in the morning, I had a working KDE installation.

But, in KDE, i realised that konsole wasn't installed . Back to the console and 'emerge konsole'. Now, I've got a konsole and an empty KDE installation. So, the rest of the morning consisted of emerging Konqueror, amarok, ymessenger, firefox and kate.

Now, it's a pretty neat installation. but a lot of things are missing. So, I decided to have the entire KDE package now, so 'emerge kde-meta'. It might take some time, but I'm in no hurry. The important things are working fine.

But, for some reason, nvidia still doesn't load up. I've tried

mknod /dev/nvidiactl c 195 255mknod /dev/nvidia0 c 195 0chown root.video /dev/nvidia*chmod 0660 /dev/nvidia*

but it still isn't working. This is the error I got:

(II) Setting vga for screen 0.(**) NVIDIA(0): Depth 16, (--) framebuffer bpp 16
(==) NVIDIA(0): RGB weight 565
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(--) NVIDIA(0): Linear framebuffer at 0xEC000000
(--) NVIDIA(0): MMIO registers at 0xFD000000
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
(EE) NVIDIA(0):  *** Aborting ***
(II) UnloadModule: "nvidia"
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found

Please consult the The X.Org Foundation support
  at https://www.x.org/wiki/
for help.
Please also check the log file at "/var/log/Xorg.0.log" for additional information.

Any suggestions?

And another thing... gentoo is definitely fast ... it starts up in 40 sec upto kdm and shutdown from KDE to power off is in 20 sec.... but I was kinda hoping for blazing speed . Of course, the sound in amarok doesn't skip no matter what I do, and the applications and programs open much faster than they did in slackware, but it's not exactly as fast as I expected it to be. I was under the impression that the windows would actually zoom around irrespective of the load.

Maybe the lag is because I haven't installed nvidia yet and only have 256MB RAM. Of course gentoo on 256MB RAM is faster than Slackware on 512MB.

I think gentoo is kinda neat. Once you have everything installed and in place, it's a pretty cool distro. I think people are afraid of gentoo (like I was) because of the huge amount of documentation to go through and the long hours to be spent in configuring the system. It loses its purpose if the system on which it is installed doesn't have a decent (atleast a 100kbps download speed) internet connection.

[qwijibow 0]

your GUI will run alot faster when you get the nvidia driver working.

run "emerge -pv nvidia-kernel"
what nvidia driver version does it report.

I had a problem with some drivers on my AMD64, it turned out to be a driver bug, and i fixed it by changing versions.

to get a list of nvidia driver versions available in portage run

ls /usr/portage/media-video/nvidia-kernel/

you can cause portage to install a different version of nvidia-driver with the command

emerge '=media-video/nvidia-kernel-<version>'

be sure that you also change nvidia-glx version to match the nvidia-kernel version.

The latest 1.0.6xxx version would probably work, this is rock stable.

but if you want to go th other way, and try a bleeding edge version which gentoo calls *probably stable but not well tested* (Meaning that nvidia have released the driver as stable, but its new to gentoo, and not tested well amungst gentoo users)

then you can install a version highter that the one gentoo installs by default.

to do this you will need to tell portage that you dont mind using *stable but not well tested* versions of nvidia kernel and lxlike so...

echo "media-video/nvidia-kernel ~x86" >> /etc/portage/package.keywordsecho "media-video/nvidia-glx ~x86" >> /etc/portage/package.keywords

(basically, add the package name, followed by ~x86 *meaning testing for x86)

then running

env-update

after you have emerged a different version of nvidia-kernel, unload the nvidia kernel like so "rmmod nvidia"

re-load it "modprobe nvidia"

set Xorg to use nvidia instead of nv

then re-start xdm.

It seems complicated, but after a while, emerge portage becomes second nature.

Gentoo's speed relies on how you install it, what CFLAGS you used, and what USE flags you set.

There are many things you can do to improve the speed further. like using native linux threads insead of the default pthreads, and using kernel's optimised for desktops, and aggressive schedulers etc etc.. but i dont want to ovrload you with Info

[abhiram 0]

Finally got nvidia to work. Apparently, there is some problem with the current nvidia drivers and kernel-2.6.13. So, I unmasked nvidia-glx, nvidia-kernel and 2 other dependencies that were required for the drivers. After emerging, just edited the xorg.conf file as given in the handbook and restarted X. I'm using the 7676 nvidia drivers. It's working great now. Time to try out some games .

There are many things you can do to improve the speed further. like using native linux threads insead of the default pthreads, and using kernel's optimised for desktops, and aggressive schedulers etc etc.. but i dont want to ovrload you with Info tongue.gif

I'll get to them as well. After spending so much time with this, I'm not leaving any stone unturned. I'm going to learn this distro down to the basics . Thanks for your help.