I have a small fixed ip network at home, running red hat 9 on two amd k6 500 Mhz boxes. One has 256 M memory and the other 320 M. They pretty much meet my needs, but lately I have detected that the internet sites I frequent are requiring some more modern software than I can run. I attempted to install Fedora 8, but that failed on the AMD processors. So I ordered a set of CentOS 5.3 i386 disks. The CentOS install fails also. I could get the first screen up, offering choices of how to boot, and if I asked for memtest86 that would start. (I had verified memory earlier, during my F8 attempts). However, any other choice resulted in a reboot (generally during loading of vmlinux). Sometimes the disk wouldn't be recognized as bootable. I have convinced myself that the disks are OK (see below) and that I must need either better hardware or more memory (but this is the i386 version of CentOS 5.3) or some parameter on the install that I haven't tried (and I've tried about all I have found or remember). I would appreciate any help. What I've done so far: At first I thought that the CentOS 1of6 disk must be bad (couldn't even run a mediacheck) and emailed the vendor. Then it occurred to me that I could perhaps download and burn disk 1of6 and use that to get the install started. (Up until my attempts at F8 I had not had the capability to burn a CD. While trying with F8 I bought a used "intel inside" computer with a broken XP system on it. I blew that away but F8 found problems with the hard disk and wouldn't install. I managed to install RedHat 9, which gave me access to the CD). I reaized that the process would be a bit "iffy" on a box with hard-disk problems, and also I had never burned a cd, but googled around for instructions and plunged in. I downloaded an ios, checked it with md5sum and it was OK. I copied it to the "new" computer via NFS and checked it again: OK. I burned a CD using cdrecord, and the burn appeared to work. The result behaved much like the original had. I tried two more times, varying stuff that I thought might affect the burn, and always got the same sort of behavior. Finally I tried mounting (on a RH9 box) each of the four disks 1of6 I now had and copied (from /dev/cdrom rather than /mnt/cdrom, so as to avoid separating out the files) each to a separate directory and compared them. All three that I burned were identical. The "store bought" disk was a little larger, but compared OK up to EOF on the other (and I recall reading that mass-produced disks might be different in their padding). So I am convinced now that there is nothing wrong with either the original or recently burned disks 1of6 and the problem must either be requiring better/more hardware (but this is the i386 version of cent OS) or some parameter on the install I have never heard of. Again, any advice would be appreciated. Buz
On Mon, Sep 14, 2009 at 3:10 PM, Buz Davis <buzdavis at earthlink.net> wrote:> > I have a small fixed ip network at home, running red hat 9 on two > amd k6 500 Mhz boxes. ?One has 256 M memory and the other 320 M. ?They > pretty much meet my needs, but lately I have detected that the internet > sites > I frequent are requiring some more modern software than I can run. ? I > attempted to install > Fedora 8, but that failed on the AMD processors. ?So I ordered a set of > CentOS 5.3 i386 disks.You need the i586 kernel for the AMD K6 machines. Bad news is CentOS 5 does not support it. Good news is CentOS 4 supports it. Visit http://i586.centos.org/centos/4/ and you will find what is required to boot your system. Once booted, type 'linux i586' to start the installation. Good luck! Akemi
On Mon, 14 Sep 2009 Akemi Yagi wrote> You need the i586 kernel for the AMD K6 machines. Bad news is CentOS 5 > does not support it. Good news is CentOS 4 supports it. > > Visit http://i586.centos.org/centos/4/ and you will find what is > required to boot your system. Once booted, type 'linux i586' to start > the installation. > > Good luck!Akemi, Thanks for the information. I did visit the link and found what I interpreted as a replacement iso for the first disk of 4. Did I misinterpret something ? I downloaded disks 2 through 4 from the Ga. Tech mirror and burned them as well, all apparently without error. However, when I attempted to run a media check from "i586 text nousb" the first disk passed but the others resulted in "unable to find install image on /tmp/cdrom" (or something close to that, I am quoting from memory). I looked at the disks and they appear to be OK, and I burned them with the same commands I used for disk 1. Does this mean I needed to gets disks 2 through 4 elsewhere ? Only disk 1 of the set was on the link I visited. Another question, and perhaps I need to google this more but thought I'd ask: On the system I am trying to install this I already have on hda msdos and RH9. I have a large disk on hdc, however, with lots of free space. It appeared that by telling the installation program to use only hdc I could successfully put CentOS there, but when it came time to tell grub what other systems to boot I it recognized dos on hda1, but I could find no way to add RH9 on hda2. It seemed that I was limited to only two operating systems. Is there a way around this ? (If not I intend to physically swap the hard disks and just install on the new hda, keeping stuff on the old hdc for reference). But if it is possible to keep the two existing systems on hda and put CentOS on hdc, and have grub on the mbr of hda control the boot I'd like to do that. Thanks, Buz Davis
First, let me apologize to Akemi Yagi for my delay in answering his last query. After failing with 4.8 I realized that I had at least one and possibly two computers capable of handling 5.3 (one was newly purchased second-hand and had many unknowns about it. The other was well documented but its power was obscured by a thick layer of Windows XP). To answer Akemi's last questions, yes, I am sure that the disks were all 4.8. I checked with md5sum before burning the cds and am not sure how to check further than that. I never got past using the first (modified) disk that I downloaded from the url you gave me, because anaconda failed to find a file - having to do with GNOME, so the process aborted. Again I apologize for delaying my response, but I have been occupied again with 5.3, but on the newer computers. 5.3 installed easily and correctly on the first machine (celeron, 1.2Ghz, 512M ram, 80G disk). With the second machine I am getting consistent failure to load X, despite many attempts that appeared to be good installs. For the last attempt I had about 160G of disk available, over 800M of ran, and (as reported by the BIOS, a pentium 3 processor. Unfortunately a remaining big unknown for that system is the Mhz rating for the processor. When I got it, with a broken XP system on it, it was set at 550Mhz. The BIOS allows three choices: 366, 500, and 7xx (I don't remember the last value precisely, but the system wont boot at all, so it's moot). I have mainly messed with the 550 setting, but have occasionally tried at 360. It doesn't seem to make a difference. On that system the output of uname -a gives 2.6.18-128.3l5 for the kernel, and "i686 i686 i386" at the end. Free shows 905240 total and 231688 used. Console mouse services work. When I attempt to load X I get the hollow-x cursor, then the black screen with the nautilus window and then white bars at top and bottom, along with an arrow cursor. The white bars get populated with some icons and then comes the blue screen with the doily-like design, but at that point the system appears to be hung. The computer, home, and trash icons do not appear on the desktop, nor does a mouse-pointer, and wiggling the mouse does nothing. I have compared the /var/log/Xorg.0.log files from the two machines (one that works, one that doesn't), and they both end in the same place: <snip> (--) <default pointer>: PnP-detected protocol: "ExplorerPS/2" (II) <default pointer>: ps2EnableDataReporting: succeeded <eof> I would appreciate any help in getting X to work on this second computer Thanks, Buz Davis
> > Eric Clark wrote: > have you tried getting KDE to run? >Eric, Thanks, that's a good suggestion to try KDE - so far I haven't figured out how to install it short of a complete reinstall. If I run yom I get a message to the effect that a database is malformed. It downs't say if the database is mine or the morrors, though. What should I do about that ? I hould have mentioned that I ran memtest86 after installing some new memory (this was before my last install of CentOS). I let it complete one pass (about an hour and a half) and it found no errors. Thanks, Buz Davis
I am at a lost with this server of mine. I have replaced everything except just replacing the entire server. The server is running at a low load but every so often it starts to have high packet loss / latency(the average ms for me is 80 but it jumps to 4000 during this period) and eventually becomes unresponsive. All traffic seems normal, no unusual activity. But once or twice a month it starts to get packet loss and after an hour or two it crashes. The network itself is fine. All other servers on the same subnet work fine. Any ideas what could be causing this kind of behavior? Thanks, Tim