cobalt qube
by
Anonymous Coward
·
· Score: 4, Interesting
i run a modified version of redhat 4.2 on a cobalt qube. It's 150 mhz and it has 32 mb ram, and works as a great fileserver for 10 users. whoever said you need alot of computing power for a server is wrong
Re:cobalt qube
by
danamania
·
· Score: 5, Interesting
33Mhz, 32Mb and a 250mb HD for my debian web server. It's served about 320mb in 24 hours (across a slow-arsed outbound link unfortunately) not long ago and took things in its stride. RAM usage hovers between 8 and 15mb.
Of course, I do go and post links to it here don't I:).
Re:cobalt qube
by
sys$manager
·
· Score: 5, Interesting
whoever said you need alot of computing power for a server is wrong
That's a pretty broad generalization to make. You may not need a lot of computing power for a 10 user file server (and anyone who says you do is a total moron) but there are applications for which you do need a lot of power.
I see the mods fell for your troll though.
Re:cobalt qube
by
Chromium_One
·
· Score: 4, Interesting
How about for a 4-person fileserver?
Had an AMD 386DX40 box, 20MB RAM, 40MB HDD running Slackware 8. Managed to strip the install down to a hair under 20MB with a few system tools, Apache, Sendmail, and a couple userland goodies like PINE. It was running as households' gateway/firewall box to a 512k DSL hookup. Worked pretty well... max uptime was about 6 months (I was out of state for most of that time).
Eventually added another harddrive and turned it into the household MP3 server as well. Worked fine most of the time =)
-- When you live in a sick society, just about everything you do is wrong.
PDA anyone?
by
Slashdotess
·
· Score: 5, Interesting
I run a light version of linux on my Compaq Ipaq and I think it's great. I can't wait until it becomes good enough to go into pda's full scale and replace proprietary OS's like palm.
Re:Older OS's?!?!
by
AndroidCat
·
· Score: 5, Interesting
What about networking? Most MS-DOS networking was really butt-ugly.
I'm currently going for a FreeBSD install on an older machine because it has an easy network-bootstrap install.
I did shoehorn Win98 onto a 486/66 for my burglar alarm, but it's not a pretty sight.
-- One line blog. I hear that they're called Twitters now.
Can someone with more knowledge give me some more info on the differences between DietLibC and uclibc? As in how much I save in binary size for both of them. Problems (something like it wont support translations is a big thing) such as feature Y wont work. Can I compile Gnome or KDE with them? I read the FAQ and both seem wonderfull and I really don't see why someone _wouldn't_ want to use them. So why wouldn't I want to use them?
-Benjamin Meyer
-- Do you changes clothes while making the "chee-chee-cha-cha-choh" transformation sound?
Re:Comparison
by
captaineo
·
· Score: 4, Interesting
First you have to understand that the main design goal of glibc is code bloat. (I'm not kidding, static hello-world.c is >100KB). A lot of this is because glibc tries to support a large API, most of which is never used by most programs (e.g. locales). Also, glibc tends to lump lots of stuff into individual object files, so the linker does a worse job of discarding unused code. (the first call to printf() in your program pulls in tens of KB of stdio code)
The concept of uclibc and dietlibc is to support the 10% of APIs that 90% of programs use, and to behave better with static linking. uclibc makes some sacrifices to work better with glibc-based software (I don't have exact numbers since I don't use uclibc). dietlibc goes completely to the extreme; anything that can be cut is cut, and object code is carefully divided so that static executables only include the code they really need.
For concrete examples, see the static binaries compiled by dietlibc's author at ftp://foobar.math.fu-berlin.de:2121/pub/dietlibc/b in-i386 (cat is 3KB, tar is 63KB, the thttpd web server is 42KB - add them up and you are just about equal to hello-world.c with glibc). Compare these with the sizes of even dynamically-linked glibc binaries on your own system.
The reason you wouldn't want to use a cut-down libc for something like Gnome or KDE is that you'd have to recompile your entire system, including the X libraries and all other dependencies. (you can't use your existing X libraries since they are already linked to glibc). Along the way you are sure to run into one or two obscure C library APIs that only glibc implements.
I think eventually there is a chance that glibc will be replaced by one of the cut-down libcs. The degree of bloat in glibc is simply obscene, and on top of that there is the backwards-incompatibility problem. (many packages broke during the transitions from 2.1 to 2.2 to 2.3, which should never have happened with a stable thing like the C library). Linus himself has even floated this suggestion on the LKML. The question that remains is whether the full glibc API can be implemented without creating another bloated monster. (there is no real alternative, since glibc's API has been enshrined in the LSB already...)
On a similar note, I'd love to see GCC drop the problematic GNU STL for STLport. In my tests STLport has about half the cost in compile time and code size...
Glibc is optimized for speed and standards compliance. It's also what Gnome and KDE and everything else on Linux is tested with, and has vastly more testers. I had 128 MB of RAM on this box when I bought it four years ago, and it wasn't top of the line. What's a half of MB of memory, especially as it cost me less than $50 to upgrade it to 384 MB?
I just ran memstat on my box. I'm running a konsole, and mozilla and emacs. Glibc is pretty far down on the list of memory wasters. Mozilla takes up 22 MB; xfs 15 MB; QT 5 MB; Emacs 3 MB; bbkeys 5 MB (I smell a memory leak); libkio 2 MB; 1.2 MB for each of libgtk and libkdecore. Deep down in this list is glibc, taking up just over 1 MB. If you're going to be running Gnome and/or KDE, glibc is not your memory waster.
First you have to understand that the main design goal of glibc is code bloat. (I'm not kidding,
But you've got to be kidding. It is a simply absurd statement. I don't think you understand the situtation, because to understand the situation you would have to know what the real design goals of glibc are and how they affected the library.
static hello-world.c is >100KB)
Okay, and how big is dynamically linked hello-world.c? There aren't that many reasons to statically link a program. There may be some on reducing program size, but I would think Emacs and OpenOffice and Mozilla - the many megabyte executables - would be much more interesting than 100KB.
most of which is never used by most programs (e.g. locales).
How do you measure that? Every program that's not a server needs to be using locales; returning localized messages and sorting information the way the user would expect it are two big things.
The degree of bloat in glibc is simply obscene, and on top of that there is the backwards-incompatibility problem.
What exactly is an obscene amount of bloat? I have QT, GTK, 3 KDE libs and 2 Mozilla libs loaded into memory, each of which is larger than glibc. Why should I worry about the 8th largest library open on my system?
The uclibc people understand that they were trading speed and standards-complance for size, and know that it's not a good tradeoff for everyone. Do you really understand what tradeoffs were made in glibc, well enough to make a better library?
The reason why there's the backward compatibility problem is two-fold; first, people keep trying to link directly to glibc's internals, and not changing those would be a pain, and second, you want to make major improvements, but changing the libc major number is a flag day, so they try to support old stuff while making major changes, with some success and some failure.
The question that remains is whether the full glibc API can be implemented without creating another bloated monster. (there is no real alternative, since glibc's API has been enshrined in the LSB already...)
There is no real alternative, because most of glibc's API comes from POSIX and Single Unix Standard (SUS), or from traditional BSD functions.
386SX16, 4M RAM, no HD, parallel port
by
czaby
·
· Score: 5, Interesting
Once I had a happy linux on a 386SX-16MHz very old laptop, without any working hard disc. The floppy was enough to boot it, 4 Megs RAM is perfect for a small kernel, some shells and telnets, everything else (even the swap) comes through PLIP on the printer port. It was much funnier than my VT420 terminal:)
Re:From the other end of the discussion...
by
fishbowl
·
· Score: 5, Interesting
Debian is already lightweight. Install the base system and whatever drivers you need. Don't select anything in dselect, and you're done; installs in under 10 megs.
-- -fb
Everything not expressly forbidden is now mandatory.
Re:From the other end of the discussion...
by
Tetsujin28
·
· Score: 4, Interesting
Maybe I'm lucky, or have low expectations, but I run Debian with GNOME on a Pentium 150. Works fine for me. I can surf the Web, read email or do word processing while listening to MP3s on XMMS. Granted, XMMS didn't work very well until I recompiled my kernel.
IBM actually has a quadruple DNS A resource record for maximum load balancing and parallelism in a Class B network spread across none less than 4/16 subnets. I kid you not:
-- "The lesson to be learned is not to take the comments on slashdot too literally." --Vinnie Falco, BearShare
Re:Mini-distros
by
Istealmymusic
·
· Score: 4, Interesting
Wait a minute...you say "no need for even a hard drive" and then "it's possible [...] to run a working mail-server on an old 386 with them". Where would you store your mail queue, on a RAM drive or NFS partition?
-- "The lesson to be learned is not to take the comments on slashdot too literally." --Vinnie Falco, BearShare
Two quick points:
by
Alethes
·
· Score: 5, Interesting
1) Using a light linux distro on a really fast machine just makes it seem that much faster. There's no need to try to find some old and slow machine to take advantage of a fast and light distro.
2) The versatility of Linux is really inspiring. We have everything from floppy distros, and game machines to Gnome, KDE and Lycoris all using variations of the same kernel. I, for one, think that's pretty cool.
FreeBSD may be an option
by
Billly+Gates
·
· Score: 5, Interesting
FreeBSD works great with minimal hardware due to the absensce of bloatware on most modern linux distro's. The Bsd daemons like inet are much less resource intensive then their Linux counterparts. By default FreeBSD only has minimal daemons running.
Gentoo may be another option due to its liteness upon default install. Everything and I mean everything must be configured and installed via "emerge x". This is also the downside. IF you have a slow 386 and a 28. modem for an internet connection you can expect s several day installation.
NetBSD seems popular with many users with old machines like ancient macs. It may be more minimalist but I have never used it. Perhaps someone who has could care to comment. I like FreeBSD because of the excellent book that comes with the box set which will be helpfull since you will not have any of the gui point and click utilities like anaconda and yast2 to setup your 386.
I like Linux myself because I am use to the SYSV init. I do not wish to start a flameware but FreeBSD is great for minimal installs and come with the best console documentation. It has its uses and if your use standard free software like sambe or apache, then a *BSD variant or Linux one wont matter.
Not Impressed
by
RAMMS+EIN
·
· Score: 4, Interesting
Sorry, I'm not impressed by this. 12 MB? I think I can run a vanilla Linux distro on that. I have and old IMB PS/ValuePoint 425SX/S (25 MHz, no FPU, 4 MB RAM that somehow refuses to be upgraded...anybody know why?) here that I wanted ti run Linux on. I couldn't get any Linux install floppy to boot and work on it, not even the ones that were advertised as working with 4MB core. I suppose this is due to IBM eating up 384 KB of memory (shadow ROM or something). Anyway, I made my own bootdisk with a 2.4.19 kernel with networking support, module for my NIC, and an uncompressed filesystem with busybox, fdisk, ip, mke2fs, insmod, mkswap and swapon, all statically linked to dietlibc. It worked great on my machine, allowed me to partition the hard drive, create and mount the swap partition, make and mount an ext2 file system, and install the files necessary for booting. I'm currently working on a uClibc-based system with picogui. It's almost finished and will be available from my website once it is. It could take some time, though, cause I'm overloaded with work these days.
--- Goto, n.:
A programming tool that exists to allow structured programmers to complain about unstructured programmers.
-- Ray Simard
It took a lot of searching and a few false starts, but I finally got Linux going on my old laptop a few years ago. I guess I did it mostly for the challenge.
Specs:
386sx @ 16Mhz 5 Megs RAM (subtract a bit for BIOS shadowing...) 240MB HDD (half DOS, half Linux ext2) No PCMCIA, Ethernet, or IR ports. Currently boots MS-DOS/Win3.1 and then uses LoadLinux. Installed: Perl, GCC tool chain, vi, and just barely enough of everything else to get by.
I tried FreeBSD first -- that's what I normally run on my Unix boxes. However, while it can run on 5MB, it is a real challenge to get it installed with only 5MB -- the installer needs 8MB, and with no swap partition set up, it can only use RAM.
I came to the conclusion that the main problem with running a nice OS on not-so-nice hardware is getting a swap partition set up. Once Linux and FreeBSD have a little virtual memory to use, they can get by on just 4MB. But until the swap partition is mounted, everything has to squeeze into that 4MB, and it simply doesn't work.
I tried a few other distros before I finally found something that worked. It was called "ZipHam Linux." It was a derivative of Slackware running 2.0.38, and specialized for HAM radio enthusiasts. Once I had a swap partition set up, I could actually do stuff. I transferred packages via MS-DOS's InterLnk (parallel cable) and upgraded to the latest kernel I thought would work.
Recompiling the kernel on a 386sx with 4MB of RAM is an exercise in patience. I think it took about 23 hours. But it compiled! Yay. And booted.
About a year later, I graduated from college, got a better job, and bought a more reasonable laptop. As a result, the old one doesn't see much use anymore. But I still think it is pretty cool. And since it is actually the only Linux box in the house (1 FreeBSD box, and I run Linux under Bochs occasionally, but no other hardware dedicated to Linux), I sometimes fire it up just for kicks.
-- Time flies like an arrow. Fruit flies like a banana.
i run a modified version of redhat 4.2 on a cobalt qube. It's 150 mhz and it has 32 mb ram, and works as a great fileserver for 10 users. whoever said you need alot of computing power for a server is wrong
I run a light version of linux on my Compaq Ipaq and I think it's great. I can't wait until it becomes good enough to go into pda's full scale and replace proprietary OS's like palm.
I'm currently going for a FreeBSD install on an older machine because it has an easy network-bootstrap install.
I did shoehorn Win98 onto a 486/66 for my burglar alarm, but it's not a pretty sight.
One line blog. I hear that they're called Twitters now.
Can someone with more knowledge give me some more info on the differences between DietLibC and uclibc? As in how much I save in binary size for both of them. Problems (something like it wont support translations is a big thing) such as feature Y wont work. Can I compile Gnome or KDE with them? I read the FAQ and both seem wonderfull and I really don't see why someone _wouldn't_ want to use them. So why wouldn't I want to use them?
-Benjamin Meyer
Do you changes clothes while making the "chee-chee-cha-cha-choh" transformation sound?
Once I had a happy linux on a 386SX-16MHz very old laptop, without any working hard disc. :)
The floppy was enough to boot it, 4 Megs RAM is perfect for a small kernel, some shells and telnets, everything else (even the swap) comes through PLIP on the printer port.
It was much funnier than my VT420 terminal
Debian is already lightweight. Install the base system and whatever drivers you need. Don't select anything in dselect, and you're done; installs in under 10 megs.
-fb Everything not expressly forbidden is now mandatory.
Maybe I'm lucky, or have low expectations, but I run Debian with GNOME on a Pentium 150. Works fine for me. I can surf the Web, read email or do word processing while listening to MP3s on XMMS. Granted, XMMS didn't work very well until I recompiled my kernel.
- - - -
The real Tetsujin 28 is a giant robot.
"The lesson to be learned is not to take the comments on slashdot too literally." --Vinnie Falco, BearShare
Wait a minute...you say "no need for even a hard drive" and then "it's possible [...] to run a working mail-server on an old 386 with them". Where would you store your mail queue, on a RAM drive or NFS partition?
"The lesson to be learned is not to take the comments on slashdot too literally." --Vinnie Falco, BearShare
1) Using a light linux distro on a really fast machine just makes it seem that much faster. There's no need to try to find some old and slow machine to take advantage of a fast and light distro.
2) The versatility of Linux is really inspiring. We have everything from floppy distros, and game machines to Gnome, KDE and Lycoris all using variations of the same kernel. I, for one, think that's pretty cool.
Gentoo may be another option due to its liteness upon default install. Everything and I mean everything must be configured and installed via "emerge x". This is also the downside. IF you have a slow 386 and a 28. modem for an internet connection you can expect s several day installation.
NetBSD seems popular with many users with old machines like ancient macs. It may be more minimalist but I have never used it. Perhaps someone who has could care to comment. I like FreeBSD because of the excellent book that comes with the box set which will be helpfull since you will not have any of the gui point and click utilities like anaconda and yast2 to setup your 386.
I like Linux myself because I am use to the SYSV init. I do not wish to start a flameware but FreeBSD is great for minimal installs and come with the best console documentation. It has its uses and if your use standard free software like sambe or apache, then a *BSD variant or Linux one wont matter.
http://saveie6.com/
Sorry, I'm not impressed by this. 12 MB? I think I can run a vanilla Linux distro on that. I have and old IMB PS/ValuePoint 425SX/S (25 MHz, no FPU, 4 MB RAM that somehow refuses to be upgraded...anybody know why?) here that I wanted ti run Linux on. I couldn't get any Linux install floppy to boot and work on it, not even the ones that were advertised as working with 4MB core. I suppose this is due to IBM eating up 384 KB of memory (shadow ROM or something). Anyway, I made my own bootdisk with a 2.4.19 kernel with networking support, module for my NIC, and an uncompressed filesystem with busybox, fdisk, ip, mke2fs, insmod, mkswap and swapon, all statically linked to dietlibc. It worked great on my machine, allowed me to partition the hard drive, create and mount the swap partition, make and mount an ext2 file system, and install the files necessary for booting. I'm currently working on a uClibc-based system with picogui. It's almost finished and will be available from my website once it is. It could take some time, though, cause I'm overloaded with work these days.
---
Goto, n.:
A programming tool that exists to allow structured programmers
to complain about unstructured programmers.
-- Ray Simard
Please correct me if I got my facts wrong.
It took a lot of searching and a few false starts, but I finally got Linux going on my old laptop a few years ago. I guess I did it mostly for the challenge.
Specs:
386sx @ 16Mhz
5 Megs RAM (subtract a bit for BIOS shadowing...)
240MB HDD (half DOS, half Linux ext2)
No PCMCIA, Ethernet, or IR ports.
Currently boots MS-DOS/Win3.1 and then uses LoadLinux.
Installed: Perl, GCC tool chain, vi, and just barely enough of everything else to get by.
I tried FreeBSD first -- that's what I normally run on my Unix boxes. However, while it can run on 5MB, it is a real challenge to get it installed with only 5MB -- the installer needs 8MB, and with no swap partition set up, it can only use RAM.
I came to the conclusion that the main problem with running a nice OS on not-so-nice hardware is getting a swap partition set up. Once Linux and FreeBSD have a little virtual memory to use, they can get by on just 4MB. But until the swap partition is mounted, everything has to squeeze into that 4MB, and it simply doesn't work.
I tried a few other distros before I finally found something that worked. It was called "ZipHam Linux." It was a derivative of Slackware running 2.0.38, and specialized for HAM radio enthusiasts. Once I had a swap partition set up, I could actually do stuff. I transferred packages via MS-DOS's InterLnk (parallel cable) and upgraded to the latest kernel I thought would work.
Recompiling the kernel on a 386sx with 4MB of RAM is an exercise in patience. I think it took about 23 hours. But it compiled! Yay. And booted.
About a year later, I graduated from college, got a better job, and bought a more reasonable laptop. As a result, the old one doesn't see much use anymore. But I still think it is pretty cool. And since it is actually the only Linux box in the house (1 FreeBSD box, and I run Linux under Bochs occasionally, but no other hardware dedicated to Linux), I sometimes fire it up just for kicks.
Time flies like an arrow. Fruit flies like a banana.