View previous topic :: View next topic |
Author |
Message |
GenTimJS Guru
Joined: 03 May 2003 Posts: 406 Location: NH, USA
|
Posted: Sat Nov 08, 2003 2:02 pm Post subject: silo boot failure "the file just loaded does not appear |
|
|
just finised install by-the-book with no funky options or customization on my Ultra1
silo gives the following on boot:
Boot device: disk:a file and args:
the file just loaded does not appear to be executable
my silo.conf looks like this:
partition = 1
root = /dev/hda4
timeout = 15
image = /boot/vmlinux
label = linux
thoughts?
-GenTimJS
tim@killminusnine.net _________________ -Tim Smith |
|
Back to top |
|
|
GenTimJS Guru
Joined: 03 May 2003 Posts: 406 Location: NH, USA
|
Posted: Sat Nov 08, 2003 3:13 pm Post subject: update |
|
|
i had to re-partition /boot
note to others, ignore the install docs .... /boot (/dev/sda1) needs to start at block ZERO not block one.
now I have a different problem though, when my ultra starts to load silo it gives:
boot device: disk:a file and args:
SProgram terminated
OK (PROMPT)
... this reeks of residual lilo suckiness .. thoughts? _________________ -Tim Smith |
|
Back to top |
|
|
Weeve Retired Dev
Joined: 30 Oct 2002 Posts: 641
|
Posted: Sun Nov 09, 2003 1:19 am Post subject: |
|
|
SILO cannot find the boot partition
The sparc bootloader area is really small, so SILO has to load part of itself off of the disk. The S of SILO when booting is the bootloader, and ILO is loaded off the disk.
Make sure your silo.conf is correct, and check the nvram alias you are booting from in OBP. |
|
Back to top |
|
|
m4chine Apprentice
Joined: 12 Mar 2003 Posts: 271 Location: Ventura, CA, USA
|
Posted: Thu Nov 13, 2003 8:10 pm Post subject: |
|
|
Hey guys, I get the same error as you first reported. "the file just loaded does not appear to be executable". My /dev/sda1 (my boot device) starts at block 0, not block 1. Just to confirm, this is the correct setup? I have this same error on 2 ultra's and it is driving me insane. Both have fresh installs of gentoo but cant boot to them. Please help. _________________ never trust a man who can count to 1023 on his fingers.
-m4chine |
|
Back to top |
|
|
pilage n00b
Joined: 16 Jul 2003 Posts: 19
|
Posted: Tue Nov 18, 2003 6:52 pm Post subject: |
|
|
I saw in another post somewhere.
at the ok prompt you can type
Code: | boot disk0:X
where X=1-9 seems any of them work. |
this worked for me to get the machine to start the boot process. of course if i had a working kernel i would be happy but....
there was also a command to make that change stick so it would just boot, i will have to try find it again.
Pilage |
|
Back to top |
|
|
vtaoe n00b
Joined: 16 Aug 2002 Posts: 25 Location: Pensacola, FL, US
|
Posted: Sun Nov 23, 2003 7:23 am Post subject: |
|
|
The ultralinux.org FAQ has some useful info on this one that helped me out.
In short, the SPARC installation instructions at gentoo.org are correct in that the boot partition should start at block 1, not block 0. What is omitted is that you must then set your boot device in PROM to be not the boot partition (e.g. /dev/hda1) but rather the Sun "whole disk" partition (/dev/hda3). This is because the boot info is stored at block 0, not block 1, and the only partition which addresses that is the "whole disk" partition, /dev/hda3, or disk:c to PROM.
In other words, from the ok prompt, try ok boot disk:c instead of ok boot disk or ok boot disk:a. You can make this the default behavior by issuing ok setenv boot-device disk:c [floppy] [cdrom] [net] in PROM.
I think this needs a little attention in the SPARC installation guide, as it was quite confusing for awhile, to me at least. (A bug report has been opened.) |
|
Back to top |
|
|
Blademan Developer
Joined: 05 Oct 2002 Posts: 116
|
Posted: Tue Nov 25, 2003 7:46 pm Post subject: |
|
|
m4chine wrote: | ...My /dev/sda1 (my boot device) starts at block 0, not block 1. Just to confirm, this is the correct setup? |
No.
Your Sun Disk Label, aka Whole disk, partition should begin at block 0.
The actual partition with data, sda1, should begin at block 1. _________________ Blademan |
|
Back to top |
|
|
Blademan Developer
Joined: 05 Oct 2002 Posts: 116
|
Posted: Tue Nov 25, 2003 7:51 pm Post subject: |
|
|
vtaoe wrote: | ...What is omitted is that you must then set your boot device in PROM to be not the boot partition (e.g. /dev/hda1) but rather the Sun "whole disk" partition (/dev/hda3). This is because the boot info is stored at block 0, not block 1, and the only partition which addresses that is the "whole disk" partition, /dev/hda3, or disk:c to PROM.
In other words, from the ok prompt, try ok boot disk:c instead of ok boot disk or ok boot disk:a. You can make this the default behavior by issuing ok setenv boot-device disk:c [floppy] [cdrom] [net] in PROM... |
Has this been verified? The "by-the-book" install should work, since the PROM will address the drive and find the S (in SILO) to boot. I have run into having to specify for example, but that's usually a silo.conf error. _________________ Blademan |
|
Back to top |
|
|
m4chine Apprentice
Joined: 12 Mar 2003 Posts: 271 Location: Ventura, CA, USA
|
Posted: Tue Nov 25, 2003 8:14 pm Post subject: |
|
|
Blademan--
I changed my starting block, and i get the same error. the thread is https://forums.gentoo.org/viewtopic.php?t=66648&start=25 can you have a look? thanks. _________________ never trust a man who can count to 1023 on his fingers.
-m4chine |
|
Back to top |
|
|
vtaoe n00b
Joined: 16 Aug 2002 Posts: 25 Location: Pensacola, FL, US
|
Posted: Tue Nov 25, 2003 9:09 pm Post subject: |
|
|
Blademan wrote: | Has this been verified? The "by-the-book" install should work, since the PROM will address the drive and find the S (in SILO) to boot. I have run into having to specify for example, but that's usually a silo.conf error. |
It's been verified by me, anyway!
After installing Gentoo "by-the-book" I received that very error. It never even got to the "S" in "SILO" because it couldn't find the boot sector, which # silo -C /boot/silo.conf had written to block 0. In PROM, ok boot disk was causing it to look at the first block of the first partition: block 1 on the disk, since /dev/hda1 started at block 1 as per the partitioning instructions in the Gentoo-Sparc install guide. It wasn't until I told it to ok boot disk:c that it correctly tried to boot from block 0 and entered SILO. And then my silo.conf typo (image = boot/vmlinux instead of /boot/vmlinux) came into play! But that was easily corrected by simply typing boot: 1/boot/vmlinux root=/dev/hda4 at SILO's prompt.
This behavior is also documented in the UltraLinux FAQ, which is where I found the solution.
For peer review, this is my partition scheme (from fdisk):
Code: | Disk /dev/hda (Sun disk label): 16 heads, 63 sectors, 17660 cylinders
Units = cylinders of 1008 * 512 bytes
Device Flag Start End Blocks Id System
/dev/hda1 1 125 62496 83 Linux native (mounted as /boot)
/dev/hda2 u 126 1118 499968 82 Linux swap
/dev/hda3 0 17660 8900640 5 Whole disk
/dev/hda4 u 1119 17660 8336664 83 Linux native (mounted as /) |
and here is (the non-commented part of) /boot/silo.conf:
Code: | partition = 1
root = /dev/hda4
timeout = 15
default = linux
image = /boot/vmlinux
label = linux |
For reference, I am running an Ultra 5 with the stock Sun/Segate 9.1 GB IDE drive as primary master (/pci@1f,0/pci@1,1/ide@3/disk@0,0 aliased to "disk" and "disk1") and OBP 3.31.0. |
|
Back to top |
|
|
pilage n00b
Joined: 16 Jul 2003 Posts: 19
|
Posted: Wed Dec 10, 2003 9:40 pm Post subject: |
|
|
After my post i found another post(and no i don't know where it is now either it is in these forums though =) ) anyway it gave a couple of commands that helped me out.
Vtaoe goes into much more detail then i even know.
i received that error and was able to do a boot0:# (#=1-9) and it would boot from there. a few people suggested doing an envupdate to change the boot command.
i used printenv to list all the env settings and i found boot_device (i think...you will recognise it) was set to disk:a and i believe SILO is looking for a more descriptive device(hence why boot disk0 works) so i did an envupdate to change the device set in env to disk0:1. i thought this was a little cleaner and seemed to make more sense to me.
just my 2 cents
Pilage |
|
Back to top |
|
|
m4chine Apprentice
Joined: 12 Mar 2003 Posts: 271 Location: Ventura, CA, USA
|
Posted: Tue Dec 16, 2003 11:32 pm Post subject: |
|
|
unfortunately ive already tried all possible boot commands, i know the partition is correct that im trying to boot. I have a feeling its just not finding SILO for whatever reason, ill post some data tomorrow after i do more testing. all help is welcomed. _________________ never trust a man who can count to 1023 on his fingers.
-m4chine |
|
Back to top |
|
|
Raide n00b
Joined: 13 Aug 2003 Posts: 27 Location: State College, PA
|
Posted: Tue Apr 06, 2004 6:21 pm Post subject: |
|
|
i am having a similar problem booting (i think it's since an update of silo was emerged).
I've tried all of the boot commands mentioned here, but each time I get a message similar to this:
Code: |
Boot device: /pci@1f,4000/scsi@3/disk@0,0:a File and args:
SProgram terminated
ok
|
I used a boot cd and examined the /boot/silo.conf and couldn't see anything wrong. Does anyone have suggestions as to how I can boot again? i believe that /boot/silo.conf looks the same as /etc/silo.conf.. which one is important during the boot process?? I had been able to boot up fine when I first was setting up this system, but that was a while ago. I can't see what else would have caused issues besides that update of silo.
(i know I never should have powered down in the first place.. )
Edit: after some more poking around the forum i tried running
Code: | silo -C /mnt/gentoo/boot/silo.conf |
but got the error:
Code: | Fatal error: Config file /mnt/gentoo/boot/silo.conf has to be on the /dev/sr0 device |
which when mounted is read-only (i believe it's the cdrom, it contains the stages directory for example).
still confused. i think my next step is going to be to backup my XF86config file and start from scratch. only problem with that is the sparc install handbook is currently out of commision... |
|
Back to top |
|
|
Weeve Retired Dev
Joined: 30 Oct 2002 Posts: 641
|
Posted: Wed Apr 07, 2004 2:14 am Post subject: |
|
|
What you want to do is boot from the CD, mount your drives like you would for the install (dont' format them or anything, just mount). Then follow all the steps to chroot into the filesystem, but before you run the chroot command, run the following command
Code: | mount -t devfs none /mnt/gentoo/dev |
Then, once inside the chroot, run the silo -C /boot/silo.conf. Once that's done you can umount all of the drives and reboot. The system should come up again like normal. |
|
Back to top |
|
|
Raide n00b
Joined: 13 Aug 2003 Posts: 27 Location: State College, PA
|
Posted: Wed Apr 07, 2004 6:08 pm Post subject: |
|
|
thanks Weeve, that worked!
What was the problem?? Should I mount /boot and run that silo -C /boot/silo.conf every time there is a silo emerge??
thanks! |
|
Back to top |
|
|
Weeve Retired Dev
Joined: 30 Oct 2002 Posts: 641
|
Posted: Thu Apr 08, 2004 3:20 am Post subject: |
|
|
Yeah, every time silo gets upgraded, it'll update all of the .b files in /boot. When these change, silo needs to be re-run to correctly work with them. I think thers is a warning at the end of the silo emerge now to tell you do run that, but if there isn't let me know and I'll make sure to add it. |
|
Back to top |
|
|
|