1.0.1 gives modprobe errno=8

[Fenix*NBK*]

Hi all !

I have a very serious problem: Puppy Linux 1.0.1 crashes on boot (I think it is on init stage)
it writes down in text-mode:

Kmod: failed to exec /sbin/modprobe -s -k binfmt-0000, errno=8

This command is looped forever, writing as many lines on screen as it can, so I don't know what was before.

My Hardware:
AMD Athlon XP 1800+
512 MB of DDR SDRAM
VIA KM266 chipset (VIA AC97 sound + VIA Rhine II network chip)
nVidia GeForce FX 5200 128MB AGP
WD2000JB, 200GB Hard Disk
DVD+-RW drive
USB mouse
PS/2 Keyboard

maybe Puppy can't probe big hard disks ??? (>128GB)
Any ideas?
How to disable "modprobe" step ?

Please Help !!!!
P.S. All other GNU/Linux distros work fine on my machine.
I Tested: Mandrake Linux 10.1, Knoppix 3.8

[van_fanel]

[quote="Fenix*NBK*"]
> Kmod: failed to exec /sbin/modprobe -s -k binfmt-0000, errno=8

I have the same problem, but with an OLD (very old... six years old!) machine, and I guess it may be a lack of memory (surely it's not caused by a big disk, because my disk is very small!).
I'm trying Puppy 1.0.0 on this Laptop: Acer Travelmate 512T

Processor: Intel Celeron 366MHz
Memory: 32 MB RAM (!!!)
Chipset: ALi Aladdin IV 1531/1533
Video Card: NeoMagic NM2200C V.DH (NMG5) 2 MB
CD-ROM Drive 24x CD TEAC CD-224E-A26
Sound card: ESS Solo-1 E (ES1946)
PCCard : O2Micro OZ6833T Cardbus controller
Synaptic Touchpad
Modem: Lucent 56K
HD 4,1 GB (only one 1.8 GB FAT32 partition with Windows 98, the rest of the disk is free and not partitioned)

A friend of mine tested Puppy 1.0.0 under Microsoft Virtual PC: he simulated two Windows 98 machines, one with 32 MB RAM and the other with 64 MB RAM. On the 64 MB RAM machine, Puppy works correctly.
On the 32 MB machine, he gets the same "modprobe" error .

May a swap partition solve the problem? How can I create a swap partition in the free (non-partitioned) space of my hard disk? I apologize but I'm a Linux newbie!
Thank you very much in advance
Van Fanel

[PeterSieg]

Hi,

1. make a partition with cfdisk or fdisk. size app. 1.5 times physical ram. type must be
linux/swap (is 82 or 83, I always forgot ).

2. probably need to reboot.. then (in an rxvt terminal) issue the command:
mkswap <partition> ex. mkswap /dev/hda2 etc.

3. then issue: swapon <partition>

The last will be done automatically at a reboot...

PS

[Fenix*NBK*]

I forgot to mention that I already has a swap in partition /dev/hdb12
It is 2-and-a-half gigs size. +512 MB of DDR SDRAM results in 3 GB of virtual memory !!!

I don't think other users have more...
I think it proves that in my case not the memory is the bottleneck.

================
That's very strange - our machines are TOTALLY different in architecture/CPU/chipset/whatever
still has the same problem..... mmm.... what could that be ?

Is there a way to shutdown that modprobe functionality?

[danleff]

Maybe two different problems in each case? Puppy has no problem with my 200 gig drive using the Live CD, but it has a fat32 partition that Puppy uses for the pup001 file. If you just have NTFS?

Secondly, in small ram systems, as mentioned, you need swap space to make enough "memory" or swap space available to properly boot Puppy totally in memory? 64 mb may be enough, but 32 mb is not?

[van_fanel]

> Secondly, in small ram systems, as mentioned, you need swap space to make enough "memory" or swap space available to properly boot Puppy totally in memory

How can I create a swap partition if I can't even boot? I'd like to boot a "minimal" system and use it to create the partition... is there a way?
Are there some boot options I can use?
Thank you!
Van

[GuestToo]

this might be useful: http://paud.sourceforge.net/

it has parted, mkswap

you can usually make a bootable cd or a bootable floppy from a floppy image file

[Guest]

So, in my case what could be done ?
How can we debug that problem, and possibly fix in next puppy ?

-Fenix*NBK*

[Fenix*NBK*]

Hello !

I have tried Puppy 1.0.3 now, and I'm experiencing the same problem described above !!!

plz help me !!!

-Fenix*NBK*

[Bruce B]

I had the same problem when I attached another hard disk under this scenerio:

Using version 1.03

IDE 0 - master = hard drive
IDE 0 - slave = empty
IDE 1 - master = cdrom

The above works fine, but when I attached a hard disk to IDE 1 slave, it happened to me. However, it worked fine with MS-DOS

[Fenix*NBK*]

yep, I forgot to mention, that my hard disk is "hdb" - IDE0/Slave.

[GuestToo]

errno=8 usually means that the kernel can't get modules to load properly

this can happen if you have an older cpu, like an i386 trying to execute i686 instructions

this can happen if you do not have enough ram

this can happen if your ram is not working properly

in your case, probably the data being read from the cd is wrong

files are often corrupted when you download them ... you should always check the md5sum to be sure it downloaded properly ... you should have downloaded the md5sum file "puppy-1.0.3-mozilla.iso.txt" and to test the integrity of the iso file, you would type:
md5sum -c puppy-1.0.3-mozilla.iso.txt

if the iso file is ok, the cd may not have burned properly ... you can check the md5sum of the cd that same way as the iso ... type:
md5sum < /dev/cdrom
if /dev/cdrom does not work, you may have to use /dev/hdc or whatever device your cd drive actually is
if the md5sum is the same as the md5sum in the txt file, then the cd burned ok
if it is not the same md5sum number, the cd still may be ok (there may be padding characters on the cd that are changing the md5sum, or it might be the read-ahead bug)

if md5sum < /dev/cdrom does not work, you can try using dd to read the exact number of bytes ... to find the number of blocks written to the cd, type:
isoinfo -d -i puppy-1.0.3-mozilla.iso
for Puppy 1.0.3, it is:
Volume size is: 30937
use that number and type this to check the md5sum of the cd:
dd if=/dev/cdrom bs=2048 count=30937 | md5sum
the md5sum should be identical to the md5sum of the iso ... if it is, the cd burned ok ... if it is not, the cd did not burn properly, and this probably is the reason it will not boot

if the cd is ok, it probably will work ok on other machines ... if it does boot ok on other machines, but does not boot properly on your machine, then your computer is probably not reading data from the cd drive properly when it boots ... it might be the jumper settings on your drives, for example

[Ian]

Another possibility could be the age of the burner you are using, try a different burner and see if it makes a difference.

[rpossemo]

Same MODPROBE problem here. Anybody found the solution ? I don't think it's ram size, etc. Thanx !

*RPM

[danleff]

Possible Issues;

1. How much RAM is in the system? 64 mb or more? Is the CPU at least a Pentium 1?

In some laptops, ram is borrowed by some systems for video. So don't discount a shortage of ram as one problem. 4-8 mb of ram may be going to the video card.

2. A good download. As GuestToo noted, make sure you got a good download, especially on dialup, which tends to bog down with iso downloads...then you get a bad md5sum when you check it.

3. Burn the iso at a slooowww speed. 1X to 4X. Puppy is small enough to do this without taking too long.

4. I found the Puppy 1.0.3 download had to be burned very slow. Especially on a cdrw disk. It seems that the new 1.0.4 version works better. If you use a cdrw disk that had a previous version burned on it, make sure that you do a complete erase of the cdrw first, not a quick erase.

5. If you burned a cdrw on your burner and attempt to boot Puppy (or any other burned iso file image) on an old cdrom drive, it can fail with this problem. Older cdrom drives don't like burned iso images, especially on cdrw disks. Laptop cdrom drives are problematic this way. Especially older 20X or 24X speed cdrom drives on these laptops.

[Guest]

I had the same problem when I attached another hard disk under this scenerio:

Using version 1.03

IDE 0 - master = hard drive
IDE 0 - slave = empty
IDE 1 - master = cdrom

The above works fine, but when I attached a hard disk to IDE 1 slave, it happened to me. However, it worked fine with MS-DOS

I have the same problem as descibed above in version 102+103+104.
Starting Live-CD and booting Puppy results in a crash. Under 102 one of millons of errors running over the screen was:

"cp: unable to close '/new_root/root0/.etc/pcmcia/cis/COMpad2.def': No space left on device "

Puppy seems not possible to boot from CD, when there is installed only one HD as hdb slave. Also testet with another hard-disk, same result.
On the other hand: Puppy installed on a partition on the same HD (hdb slave) boots without any problems.
Any idea, what to do? Or where to read about this problem? Thanks for any help.

Gerd

[Guest]

The subject helped me to boot 1.0.4 by folowing GuestToo Puppy Master recommendation and others pointing out that the CDw/r should be fully erased before burning.

Thanks

[Ian]

Do these problems relate to laptops.

[raffy]

Your solution could be as simple as making sure that there is a hard disk at primary master (/dev/hda), and that there is a viable partition 1, either FAT/DOS or ext2/Linux (/dev/hda1).

I have tried non-primary-master hard drive before and failed.

See the link below for more details in setting-up.