Ubuntu Linux - a superb new distribution

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
A lot of people on Ars seems to be installing and enjoying it, the majority of them are Debian users and big Linux on the desktop proponents so it's a big surprise that they'd like Debian with some more ease of use but I'll be sticking with normal Debian sid. Everything I have now works so I'm not going to change for the sake of change.

I think most linux distros just seem unfinished and kind of done half assed.

Most definately do, Debian is the only one that I've felt acted like a real system instead of just a slew of packages. Gentoo may have a similar feeling but to me it's a whole lot of time and work for no real gain, I mean even ignoring the package compilation thing (which is ludicrous) and just compare emerge and apt-cache in search times, I don't have a box handy to time but last time I tried searching via emerge it took f'ing for ever where as apt-cache search takes less than a minute.
 

pac1085

Diamond Member
Jun 27, 2000
3,456
0
76
I don't have a box handy to time but last time I tried searching via emerge it took f'ing for ever where as apt-cache search takes less than a minute.

A minute?? Thats an awfuly long time. An emerge search for me takes no more than 15 seconds, 20 TOPS.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
A minute?? Thats an awfuly long time. An emerge search for me takes no more than 15 seconds, 20 TOPS.

The box I ran it on was older and had really slow disks, but even so 15-20s is terrible compared to .5s for apt-cache.
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
~ >: time apt-cache search kde
stuf....
stufff.....

real 0m1.246s
user 0m0.169s
sys 0m0.034s


~ >: time apt-cache search gnome > /dev/null

real 0m0.219s
user 0m0.177s
sys 0m0.032s


(although It doesn't bother me that emerge is slower. It's effectively a non-issue to me.)
 

xcript

Diamond Member
Apr 3, 2003
8,258
2
81
Originally posted by: pac1085
An emerge search for me takes no more than 15 seconds, 20 TOPS.

Tried esearch?

xcript@arkham ~ $ time emerge -s gnome >/dev/null

real 0m3.957s
user 0m3.185s
sys 0m0.244s

xcript@arkham ~ $ time esearch gnome >/dev/null

real 0m0.150s
user 0m0.140s
sys 0m0.009s

 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: xcript
Originally posted by: pac1085
An emerge search for me takes no more than 15 seconds, 20 TOPS.

Tried esearch?

xcript@arkham ~ $ time emerge -s gnome >/dev/null

real 0m3.957s
user 0m3.185s
sys 0m0.244s

xcript@arkham ~ $ time esearch gnome >/dev/null

real 0m0.150s
user 0m0.140s
sys 0m0.009s

Do that again, and try an emerge afterwards.
 

xcript

Diamond Member
Apr 3, 2003
8,258
2
81
Originally posted by: n0cmonkey
Do that again, and try an emerge afterwards.
I understand how it works and am aware of the limitations. ;)

It's just alot snappier for quick searches to see if something's available in portage, etc.

eupdatedb doesn't take long to recreate the index:

root@arkham ~ # time eupdatedb >/dev/null

real 0m45.150s
user 0m21.916s
sys 0m6.421s


The provided esync script calls both 'emerge --sync' and 'eupdatedb' so I just stick it in /etc/cron.daily and forget about it. :)

Obviously, though, it's usefulness would be negated by someone who sits around syncing/emerging stuff all day.
 

EmperorRob

Senior member
Mar 12, 2001
968
0
0
The only issue I've run into with Synaptic is that I haven't been able to get it to obey my /etc/apt/preferences or my /etc/apt/apt.conf file. The priorities and options there are ignored and Synaptic just has one option to select a default release.
 

user1234

Banned
Jul 11, 2004
2,428
0
0
My kernel configuration is progressing well, I'm trying to remove the drivers and modules I don't need, so the compilation will be finish faster, but there's a lot of trial and error, as sometimes I remove something essential. But I cannot get rid of the "initrd" crap without having the thing panic. Is there a way to boot without it ?
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: user1234
My kernel configuration is progressing well, I'm trying to remove the drivers and modules I don't need, so the compilation will be finish faster, but there's a lot of trial and error, as sometimes I remove something essential. But I cannot get rid of the "initrd" crap without having the thing panic. Is there a way to boot without it ?

Ahh, the beauty of custom kernels. :roll:
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Originally posted by: n0cmonkey
Originally posted by: user1234
My kernel configuration is progressing well, I'm trying to remove the drivers and modules I don't need, so the compilation will be finish faster, but there's a lot of trial and error, as sometimes I remove something essential. But I cannot get rid of the "initrd" crap without having the thing panic. Is there a way to boot without it ?

Ahh, the beauty of custom kernels. :roll:

Tis a wonderfull thing to behold.

Initrd is a loopback file system that is read off of the harddrive at boot time by the boot loader. It would contain the nessicary modules and programs needed to initialize the hardware the kernel would need to use to continue the boot up proccess.

Say you have a hardware controlled raid array, you'd need to load the drivers for that up before you access the harddrive directly, but you need the modules from the harddrive in order to get to the stupid thing in the first place. So initrd was created to solve this "the chicken and the egg" problem.

It makes making large universal style distros easier when you want a modularlized kernel with all the functionality you expect your users to need without all the unnessisary bloat and avoid any conflicting featues.

So what is probably happenning is that you have your boot loader incorrectly configured with the improper kernel options or you forgot to compile in initrd or loopback file system support, or something like that.

But if you have all the functionality nessicary to boot up and access the harddrive and it's filesystem built into the kernel and not as modules, then you don't have to worry about the initrd. Just comment it out and have root=/dev/hd** equal to the correct partition in your bootloader's configuration file.

edit:
BTW the only time I bother with custom kernels anymore is when I need to use a kernel patch for some feature or driver that isn't located in the vanilla kernel proper.
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: drag
Originally posted by: n0cmonkey
Originally posted by: user1234
My kernel configuration is progressing well, I'm trying to remove the drivers and modules I don't need, so the compilation will be finish faster, but there's a lot of trial and error, as sometimes I remove something essential. But I cannot get rid of the "initrd" crap without having the thing panic. Is there a way to boot without it ?

Ahh, the beauty of custom kernels. :roll:

Tis a wonderfull thing to behold.

Not really, IMO.
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Whatever. Kernel compiling is a not a big deal either way.

Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Also when your a newbie, it gives you a chance to screw around and learn how your system works. For many people it's the first real way they've been able to fundamentally modify some low-level behavior of the OS and make some sort of impact on it's functioning.

It's like putting headers on a old chevy with a rebuilt motor when your a teenager.
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: drag
Whatever. Kernel compiling is a not a big deal either way.

Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Also when your a newbie, it gives you a chance to screw around and learn how your system works. For many people it's the first real way they've been able to fundamentally modify some low-level behavior of the OS and make some sort of impact on it's functioning.

It's like putting headers on a old chevy with a rebuilt motor when your a teenager.

I can't say I'd enjoy doing that much either. :p

I've become extremely bored with computers all of a sudden. It's a shame really, but I can't find anything interesting enough for me to spend more than 5 minutes on it.

Anyhow, kernel compilling (when not testing or using alterate patch sets) is mostly a waste of time. It doesn't add anything, and makes support that much tougher. Just my opinion though. ;)
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Originally posted by: n0cmonkey
Originally posted by: drag
Whatever. Kernel compiling is a not a big deal either way.

Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Also when your a newbie, it gives you a chance to screw around and learn how your system works. For many people it's the first real way they've been able to fundamentally modify some low-level behavior of the OS and make some sort of impact on it's functioning.

It's like putting headers on a old chevy with a rebuilt motor when your a teenager.

I can't say I'd enjoy doing that much either. :p

I've become extremely bored with computers all of a sudden. It's a shame really, but I can't find anything interesting enough for me to spend more than 5 minutes on it.

Anyhow, kernel compilling (when not testing or using alterate patch sets) is mostly a waste of time. It doesn't add anything, and makes support that much tougher. Just my opinion though. ;)

Of course.

Computers and operating systems, generally suck.

For example: My laptop is fairly slow computer. It's cpu is fast at 2.0ghz, but it's wasted on the 4500rpm harddrive and crappy chipset. I can set the cpu at 1.2ghz with NO noticable impact on performance, unless I am compiling something and that's pretty rare.

But even then my laptop can spank the most powerfull mainframe computers sold just 5-6 years ago.

But were is the realy cool stuff that's suppose to be going on?

In free software the most talented programmers and developers are stuggling to gain enough market share to get enough critical mass to ensure that the concepts of freedom of the internet and software doesn't become extinct.

In the other camp the closed source software people are involved in a mostly big circle jerk to try to steal that extra critical 5-10% of market share from each other by doing the same tired crap that they've been doing for the past 15 freaking years. Were is the innovation?

The only cool stuff going on is the commodiziation of old big iron computer stuff, but the big names are busy trying to figure out profit from it while keeping it insanely expensive and remote enough from most people that they have to depend on paying in large sums of money to said companies just to ensure that their big database will still be around next week and not in flames.

Desktop innovation has hit a brick wall. OS X was a good attempt and has kept Apple alive so far, but MS is busy trying to convince people to buy a 6 year old OS based on a 10 year old operating system core by slapping some extra software to watch TV on it. Then their next suck fest of a OS is just going to end up being WinXP remixed, full of broken promises, but enough new features to keep fanboys alive and kicking into the next decade. Oh, and I predict that WinXP-AMD64 will never exist and the only 64bit OS that you can get for the desktop from Microsoft will be "longhorn".

Linux desktop is making good progress, performance and stability outstrips WinXP most time, but not quite it's user freindly-ness. Of course the majority of WinXP's userfreindly-ness comes from the fact that people have been using it for nearly 3 years, and Windows OSes 10 years before that.

Of course Linux desktop will never happen because Oo_Org is only 95% like MS Office and not a 130% like Office like it should be. And their is no C: drive. And their is no point and click software install.

Plus I just figured out some reg hacks to speed up my WinXP boot time by 1.7 seconds!

Boring boring boring boring.
 

n0cmonkey

Elite Member
Jun 10, 2001
42,936
1
0
Originally posted by: drag
Originally posted by: n0cmonkey
Originally posted by: drag
Whatever. Kernel compiling is a not a big deal either way.

Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Also when your a newbie, it gives you a chance to screw around and learn how your system works. For many people it's the first real way they've been able to fundamentally modify some low-level behavior of the OS and make some sort of impact on it's functioning.

It's like putting headers on a old chevy with a rebuilt motor when your a teenager.

I can't say I'd enjoy doing that much either. :p

I've become extremely bored with computers all of a sudden. It's a shame really, but I can't find anything interesting enough for me to spend more than 5 minutes on it.

Anyhow, kernel compilling (when not testing or using alterate patch sets) is mostly a waste of time. It doesn't add anything, and makes support that much tougher. Just my opinion though. ;)

Of course.

Computers and operating systems, generally suck.

For example: My laptop is fairly slow computer. It's cpu is fast at 2.0ghz, but it's wasted on the 4500rpm harddrive and crappy chipset. I can set the cpu at 1.2ghz with NO noticable impact on performance, unless I am compiling something and that's pretty rare.

But even then my laptop can spank the most powerfull mainframe computers sold just 5-6 years ago.

But were is the realy cool stuff that's suppose to be going on?

In free software the most talented programmers and developers are stuggling to gain enough market share to get enough critical mass to ensure that the concepts of freedom of the internet and software doesn't become extinct.

In the other camp the closed source software people are involved in a mostly big circle jerk to try to steal that extra critical 5-10% of market share from each other by doing the same tired crap that they've been doing for the past 15 freaking years. Were is the innovation?

The only cool stuff going on is the commodiziation of old big iron computer stuff, but the big names are busy trying to figure out profit from it while keeping it insanely expensive and remote enough from most people that they have to depend on paying in large sums of money to said companies just to ensure that their big database will still be around next week and not in flames.

Desktop innovation has hit a brick wall. OS X was a good attempt and has kept Apple alive so far, but MS is busy trying to convince people to buy a 6 year old OS based on a 10 year old operating system core by slapping some extra software to watch TV on it. Then their next suck fest of a OS is just going to end up being WinXP remixed, full of broken promises, but enough new features to keep fanboys alive and kicking into the next decade. Oh, and I predict that WinXP-AMD64 will never exist and the only 64bit OS that you can get for the desktop from Microsoft will be "longhorn".

Linux desktop is making good progress, performance and stability outstrips WinXP most time, but not quite it's user freindly-ness. Of course the majority of WinXP's userfreindly-ness comes from the fact that people have been using it for nearly 3 years, and Windows OSes 10 years before that.

Of course Linux desktop will never happen because Oo_Org is only 95% like MS Office and not a 130% like Office like it should be. And their is no C: drive. And their is no point and click software install.

Plus I just figured out some reg hacks to speed up my WinXP boot time by 1.7 seconds!

Boring boring boring boring.

I'm not sure where that mini-rant came from, but as always it was a good read. ;)
There are some neat things out there, they're just not interresting my like they used to...
 

Tbirdkid

Diamond Member
Apr 16, 2002
3,758
4
81
this thing ruined my master boot record... so i wont use it until i just want to put it on a linux box...
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Usually there's so little gain that it's in the "why?" category.
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
I use a custom kernel on my desktop system. Always have.

Right now I have 2 reasons. Originally I did it to play around with 2.6 kernels, which debian didn't have a pre-compiled version for it. then I kept on using newer versions of kernel for the bug fixes and newer features/performances. Then I needed to patch the kernel to get my wireless card going, but I don't need to do that no more.

Now it's just convienent for me to do so. It makes compiling my ivtv card easy, and it's easy to compile the nvidia driver for it.

When I upgrade my kernel, once every few months or so (usually I have a somewhat decent reason to, but I forget all of them) I just copy over my old .config file. Then do a:

make clean
make oldconfig
make menuconfig
make && make install modules modules_install

Then I come back later after it's done, check my grub.conf and reboot. Download the latest IVTV sources from some guy's cvs's snapshots, compile those. Then edit my /root/ivtvmodules script to reflect the new modules install directory, and then download and install the latest nvidia drivers.

The whole proccess takes maybe 45 minutes of real time. No sweat at all. I am a multitasker, so it's not like I am sacrificing more then a couple minutes at a time to manually go thru the steps. The most time is spent double checking the menuconfig. Hell I don't even bother with cleaning up the modules. I just make most everything modules, except for the file system and basic motherboard/network drivers and let the hotplug system take care of most of the work. Maybe run kudzu or something like that. Lots of stuff I dont' need, but maybe randomly in the distant future I may screw around with.

Some people like to use -mm patches and stuff like that to get the most performance they can out of a kernel. I don't care that much, if I want more performance I buy a faster cpu.

A 70 dollar cpu upgrade (while selling off my old one for 30-40 bucks to a freind) is well worth saving the time to try to get dubious benifits from a 3-day Gentoo-style compile/testing fest.

Other people do it for the satasfaction or the learning experiance. Lots of people do it just to get the newest Linux version or if they want to try out ReiserFS 4 or something like that.

It's not something that I'd do on a production machine or something I REALY had to depend on, (if it fails on my desktop I'll just go back to a earlier known-working kernel), but it's a routeine thing nowadays for me.

I don't bother doing it for my laptop though. Pointless, the hardware is well supported (as supported as cheap hardware can get). Even the wireless stuff is detected and setup automaticly (except for the firmware). If it ain't broke I ain't going to mess with it.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
I know there are reasons for it, I need a couple of external patches for my notebook kernel. But in general there's no reason to do it as long as your hardware is supported.
 

Klixxer

Diamond Member
Apr 7, 2004
6,149
0
0
Originally posted by: Nothinman
Most of the time it's a waste of time, but it usually takes so little effort it's usually in the "why not?" catagory.

Usually there's so little gain that it's in the "why?" category.

So you can contribute more?
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
So you can contribute more?

Probably, are you curious about a specific thing? I mean why do you think it has a big beneift? The only thing you end up doing is saving disk space from all the modules you build, but the day you plug something in that you didn't think about and it just works you'll be glad you did that.
 

user1234

Banned
Jul 11, 2004
2,428
0
0
Hello ??? I'm trying to optimize performance and just recompiling with for my CPU already gives me a 10% increase on prime 95 benchmark. I agree that the rest of the configuration, including removal of unneccessary drivers will probably not help much, but still I like the idea of a lean and fast kernel. As discussed here, some distros are already optimized for perfformance, but I didn't like the ones I tried, so I plan on optimizing Ubuntu to at least the same level. And I just finished watching 2fast2furious with gmplayer. Not boring at all.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Woohoo Prime95, who cares? Do you run Prime95 all the time while you're using the machine? I sure as hell don't. 10% in Prime95 probably means 1% or less in real world usage. And honestly, if your distro doesn't have differnet kernels for each CPU available there's something wrong.

$apt-cache search kernel-image-2.6.8
kernel-image-2.6.8-1-386 - Linux kernel image for version 2.6.8 on 386.
kernel-image-2.6.8-1-686 - Linux kernel image for version 2.6.8 on PPro/Celeron/PII/PIII/PIV.
kernel-image-2.6.8-1-686-smp - Linux kernel image for version 2.6.8 on PPro/Celeron/PII/PIII/PIV SMP.
kernel-image-2.6.8-1-k7 - Linux kernel image for version 2.6.8 on AMD K7.
kernel-image-2.6.8-1-k7-smp - Linux kernel image for version 2.6.8 on AMD K7 SMP.
kernel-image-2.6.8-2-amd64-generic - Linux kernel image for version 2.6.8 on generic x86_64 systems.
kernel-image-2.6.8-2-amd64-k8 - Linux kernel image for version 2.6.8 on AMD64 systems.
kernel-image-2.6.8-2-amd64-k8-smp - Linux kernel image for version 2.6.8 on AMD K8 SMP.
kernel-image-2.6.8-2-amd64-xeon - Linux kernel image for version 2.6.8 on Intel amd64 systems.
kernel-image-2.6.8-4-amd64-generic - Linux kernel image for version 2.6.8 on generic x86_64 systems.
kernel-image-2.6.8-4-amd64-k8 - Linux kernel image for version 2.6.8 on AMD64 systems.
kernel-image-2.6.8-4-amd64-k8-smp - Linux kernel image for version 2.6.8 on AMD K8 SMP.
kernel-image-2.6.8-4-em64t-p4 - Linux kernel image for version 2.6.8 on Intel EM64T systems.
kernel-image-2.6.8-4-em64t-p4-smp - Linux kernel image for version 2.6.8 on Intel EM64T systems.
kernel-image-2.6.8-9-amd64-generic - Linux kernel image for version 2.6.8 on generic x86_64 systems
kernel-image-2.6.8-9-amd64-k8 - Linux kernel image for version 2.6.8 on AMD64 systems
kernel-image-2.6.8-9-amd64-k8-smp - Linux kernel image for version 2.6.8 on AMD64 SMP systems
kernel-image-2.6.8-9-em64t-p4 - Linux kernel image for version 2.6.8 on Intel EM64T systems
kernel-image-2.6.8-9-em64t-p4-smp - Linux kernel image for version 2.6.8 on Intel EM64T SMP systems
kernel-tree-2.6.8 - Linux kernel tree for building prepackaged Debian kernel images