While I don't have first hand experience with the Soltek board mentionned on this page, I do have first hand experience with one of it's predecessors, the Soltek SL-75KAV based on the KT133A and I'm running an AMD 1.4GHz Athlon.
I must say, I've never had the privelidge of using such a stable machine before. Both in Linux (2.4.16:) and Windows XP. I run this machine nearly 24/7 and have never once had any strangeness in either OS. All in all, I definately recommend Soltek boards - they are fairly cost effective too.
Re:Which distribution allows easy kernel upgrades?
on
Linux 2.4.16 Released
·
· Score: 1
All the distributions allow upgrading the kernel easily. You just have to treat them differently.
Slack was no problem because it keeps the kernel in/vmlinuz (which is where the kernel install puts it as well)
Redhay, and by extension, Mandrake, put kernels in/boot/kernel-version
All you have to do is make sure that the new kernel ends up where/etc/lilo.conf is looking and rerun/sbin/lilo.
The kernel is not dependant on libraries at all, so library version conflicts should not effect a kernel upgrade process.
Someone please correct me if I'm wrong, but how can an OS trash the geometry of a drive? The geometry of a drive is something fixed in hardware - a drive has 8 heads, 1024 cylinders, etc... software can't change that.
Of course there are ways of interpreting the geometry differently (BIOS HD autodetects often report different geometry than what's on the drive case) but again, this should be out of the OS's hands.
No, he's right. If your drive is old/poor quality, and has bad sectors, data may get written perfectly from the OS's viewpoint, but if that data lands on a bad sector, it's going to be damaged and the partition SHOULD be marked unclean so that an fsck can try to sort things out next time.
This isn't an oversight on the part of the kernel or it's developers, but a fault of hardware not doing it's job. This is what the author of that comment was referring to.
I got my start in Linux from Slackware as well... I don't recall off-hand what version of Slack it was (3.0 maybe?), but it was using kernel 1.2.13.
Like others here have done, I also went astray - flirting with Mandrake for short time - but I dutifully returned to Slackware after I finally got fed up with dependancies that impeded the ability to upgrade rather than assisted it.
I'm a guy that enjoys compiling from source, being able to stay on the bleeding edge of stuff with pre-alpha's just to see what the devel team have put together. I like fixing things when they don't work, but not having to fix the system because some GUI config tool screwed things up.
I certainly hope Slackware is here to stay. Big props go out to Patrick and rest of the Slackware crew - keep up the great work!
Gawd, I hope it's not harmful! The manufacturer of my PC included the unused portion of Arctic Silver II with my order. Being the curious cat I generally am, I squeezed a bit out on to my finger... damn is that hard to wipe off without managing to get it everywhere!:)
For some time a few years ago, the computer industry toyed with the idea of putting their CPU's in game cartridge-like packages (Slot 1, Slot A, etc...). Both Intel and AMD did this - however it didn't last long and finding slot processors and boards is next to impossible today.
I don't know the technical reasons behind it, but the industry switched back to socket type CPUs (almost simultaneously) and so you see Socket 370 (Intel) and Socket A (AMD) on most if not all new mainboards.
I'm running a 1.4GHz T-Bird with the CoolerMaster that was included with the system, it is thermally connected to the chip with Arctic Silver II. I also added 2 3" case fans to the system (one in front, one in back).
The sensor consistantly shows an average temp of about 39C going as high as 48C.
I just mail ordered the majority of a new Athlon 1.4GHz PC and was quite impressed to find that the company had used plastic zip-ties to secure the heatsink - and CPU - to the board (utilizing the holes in the board that are there for industrial strength cooling).
They also put zip-ties around the clips on the RAM to ensure that it didn't come loose either.
Well, if humans can let hoards of machines do their battles and spare human lives... go for it. On the other hand, if we could stop fighting altogether, we could use those machines for far better purposes.
Slashdot Mirror
I was going to write that exact same comment - more or less :)
Lets hear it for the Rio PMP300 - mine's still thumping along.
While I don't have first hand experience with the Soltek board mentionned on this page, I do have first hand experience with one of it's predecessors, the Soltek SL-75KAV based on the KT133A and I'm running an AMD 1.4GHz Athlon.
:) and Windows XP. I run this machine nearly 24/7 and have never once had any strangeness in either OS. All in all, I definately recommend Soltek boards - they are fairly cost effective too.
I must say, I've never had the privelidge of using such a stable machine before. Both in Linux (2.4.16
All the distributions allow upgrading the kernel easily. You just have to treat them differently.
/vmlinuz (which is where the kernel install puts it as well)
/boot/kernel-version
/etc/lilo.conf is looking and rerun /sbin/lilo.
Slack was no problem because it keeps the kernel in
Redhay, and by extension, Mandrake, put kernels in
All you have to do is make sure that the new kernel ends up where
The kernel is not dependant on libraries at all, so library version conflicts should not effect a kernel upgrade process.
> The real problem is that new functionality is being added to the stable branch
In this case, the real problem was that a bugfix (which is supposed to occur in stable kernels) was faulty and caused another bug.
Someone please correct me if I'm wrong, but how can an OS trash the geometry of a drive? The geometry of a drive is something fixed in hardware - a drive has 8 heads, 1024 cylinders, etc... software can't change that.
Of course there are ways of interpreting the geometry differently (BIOS HD autodetects often report different geometry than what's on the drive case) but again, this should be out of the OS's hands.
No, he's right. If your drive is old/poor quality, and has bad sectors, data may get written perfectly from the OS's viewpoint, but if that data lands on a bad sector, it's going to be damaged and the partition SHOULD be marked unclean so that an fsck can try to sort things out next time.
This isn't an oversight on the part of the kernel or it's developers, but a fault of hardware not doing it's job. This is what the author of that comment was referring to.
I got my start in Linux from Slackware as well... I don't recall off-hand what version of Slack it was (3.0 maybe?), but it was using kernel 1.2.13.
Like others here have done, I also went astray - flirting with Mandrake for short time - but I dutifully returned to Slackware after I finally got fed up with dependancies that impeded the ability to upgrade rather than assisted it.
I'm a guy that enjoys compiling from source, being able to stay on the bleeding edge of stuff with pre-alpha's just to see what the devel team have put together. I like fixing things when they don't work, but not having to fix the system because some GUI config tool screwed things up.
I certainly hope Slackware is here to stay. Big props go out to Patrick and rest of the Slackware crew - keep up the great work!
Gawd, I hope it's not harmful! The manufacturer of my PC included the unused portion of Arctic Silver II with my order. Being the curious cat I generally am, I squeezed a bit out on to my finger... damn is that hard to wipe off without managing to get it everywhere! :)
For some time a few years ago, the computer industry toyed with the idea of putting their CPU's in game cartridge-like packages (Slot 1, Slot A, etc...). Both Intel and AMD did this - however it didn't last long and finding slot processors and boards is next to impossible today.
I don't know the technical reasons behind it, but the industry switched back to socket type CPUs (almost simultaneously) and so you see Socket 370 (Intel) and Socket A (AMD) on most if not all new mainboards.
I'm running a 1.4GHz T-Bird with the CoolerMaster that was included with the system, it is thermally connected to the chip with Arctic Silver II. I also added 2 3" case fans to the system (one in front, one in back).
The sensor consistantly shows an average temp of about 39C going as high as 48C.
I just mail ordered the majority of a new Athlon 1.4GHz PC and was quite impressed to find that the company had used plastic zip-ties to secure the heatsink - and CPU - to the board (utilizing the holes in the board that are there for industrial strength cooling).
They also put zip-ties around the clips on the RAM to ensure that it didn't come loose either.
I thought that it was a very nice touch.
It's funny cause it looked as if you were poking fun at the kernel version numbering scheme to explain the 2.4.11 bug. I found it at least amusing. :)
Well, if humans can let hoards of machines do their battles and spare human lives... go for it. On the other hand, if we could stop fighting altogether, we could use those machines for far better purposes.