I have been trying to reinstall Ubuntu LT 12.04 since my original install mysteriously went kaput.
I cleared out my second hard drive for it, moved over from my SAM460ex when I quit using that. Media Toolbox recognizes it as Hitachi JP6940... Version JP4O. It is easy to partition with Media Toolbox, and easy to format, and copy files to.

Meanwhile, the Install program on the Ubuntu LT DVD recognizes that drive as /dev/sdb, and seems to create partitions on it readily. But then when I click on "Install," I consistently get an error message claiming read/write problems. OS4.1FE has no trouble writing to or reading from it, so why can't Ubuntu Install do it?

I have tried to reformat it with the Disk Utility on the Ubuntu DVD, but no luck. But that utility is trying to install to a disk it sees as Model: ATA iHathc i..., not Hitachi ... It gets the Hi, the ta, the ch, and the i[space] all backwards. Ditto Firmware Version: PJO43AAM instead of JP4O. (The same thing happens to my OS4 disk, /dev/sda -- ST1000 becomes TS0100...)

Incidentally, sometimes when I run Disk Utility on it, I get "SMART Status: [green dot] Disk is healthy." Othertimes I get "Disk not smart capable," or words to that effect.

Does anyone have any idea what is going on here, and whether it is fixable? I have asked the same question at Hyperion Forums, X1000 Linux Only forum, but so far no answer. Maybe there is no answer...

Last edited by kilaueabart on 27-Mar-2015 at 02:17 AM.

@kilaueabart

Could it be a transfer problem, missing some status bits?
Swapping like that looks like an endian issue, as if the drive have some magic code that tries to determine the endian of the machine..? Or the Linux driver is not properly ported to big-endian PPC.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@kilaueabart

Right now your HD is not MBR formatted. You have been playing with this for some time.

Perhaps you should download the Debian Network Install and burn it to a CD. Debian uses a program called Partman for working with HDs and if nothing else it may allow you to set the HD for MBR, and then install Ubuntu. I really don't think the Hitachi scramble is a problem with the utility not changing the RDB to MBR.

When you do select Fromat Drive, does it bring up a new window where you can select MBR before it pops up the error message you get?

_________________
Tinkering with computers.

@Deniil715
The Linux driver problem seems more likely, since there is no trouble with the drive in OS4. I've looked at http://wiki.scummvm.org/index.php/HOWTO-Debug-Endian-Issues, but so far I can't tell if it offers a solution.

@Signal
I'll have to run the disk again to be certain, but as I recall, I did select MBR before clicking Format and getting the error message.

I previously had Ubuntu running on this drive, since early 2014, until it suddenly quit working a month or so ago. (I put it in empty space on a disk that I continued to use for its Amiga partitions, which I have been told is a no-no.) Does the fact that I had been able to install it easily and use it for a year mean that my drive had been MBR formatted, but then got unformatted? Or Ubuntu just recently requires MBR formatting? This is all very confusing.

If that Debian Network Install will work on an X1000 I'll probably give it a try.

@Signal

Why should MBR be needed? MBR is an archaic format native to x86 hardware with hacks to extend its partition limits. And for PPC Linux would usually expect the Apple APM format to be used. Any Linux distro worth its salt should be able to use GUID now.

@Hypex

Yes, GUID should work, but also MBR will work. However, it is getting Linux to even change the partition table to MBR, much less GUID, is the problem he is having.

_________________
Tinkering with computers.

@kilaueabart

Quote:

kilaueabart wrote: This is all very confusing. If that Debian Network Install will work on an X1000 I'll probably give it a try.

No, MBR is pretty much standard. Amiga RDB is the oddball.

There is one other thing you might try.

After booting up the Live CD, try starting Synaptic Package Manager (which should be in the Administration menu) and installing gnu-fdisk (just type fdisk in the quick search box).

Then open a terminal and start fdisk with 'fdisk /dev/sdb'. Then type 'm' to get a help menu.
You will notice that 'o' should create a new DOS partition table, that can take some time, then 'w' will write the table to the HD. Wait till it finishes and then reboot.

Remember when you reboot the fdisk program will be gone, but if it worked you should not need it again.

Or, just plug the disk into a windows box and format the drive there.

_________________
Tinkering with computers.

@Signal

Interesting (but ultimately frustrating) day! Booted DVD Ubuntu and looked for Synaptic Package Manager (a name from long ago Debian days). I haven't found that yet, but when I came to GParted Partition Editor I gave it a look. It claimed I already had a bunch of partitions on the disk! Unallocated, sdb1, sdb2, sdb3, sdb4, unallocated, sdb5, unallocated. I deleted them all, then tried but failed to set up new ones with that program.

I went to Disk Utility; SMART was once again saying /dev/sdb1 was "healthy," and though it still had the diskname all mixed up, I looked at SMART data and ran self-tests successfully. "Format Drive" very quickly reported it had formatted 1.0 TB Free Space as Master Boot Record, and let me make a 20 MB ext4 partition as /sdb1 in it.

At this point I went back to the Install program, which no longer reported an error when I asked it to install in /dev/sdb1. For quite a long time things seemed to progress smoothly, until I got a notice that something had failed, and suggested I send data which the program would compile for me to a debugging outfit somewhere. I agreed and supposedly somebody will get back to me about the failure. Meanwhile, it still seemed to be installing so I just went ahead with the procedure until it was apparently done.

I was optimistic that it would run from CFE. The attempt took quite a while because I simply cannot type one of those "setenv -p MENU_#_COMMAND ..." things without one typo, a different one each time I redo it. But I finally got a command the machine was willing to try, until soon it told me:

Loader:elf Filesystem:raw Dev:cf0 File:vmlinux-3.18 Options:(null)
Loading: Failed.
Could not load cf0:vmlinux-3.18: Not an ELF-format executable
*** command status = -11

As I was typing those lines I vaguely remembered having intermittent trouble with kernel 3.18 before, and I think I may have time to give 3.14 a try tonight, as long as I don't make too many typos redoing MENU_3_COMMAND.

Drat! kernel 3.14 isn't "ELF-format executable" either. I've still got 3.12, 3.17, and 3.19 to try, but not tonight. I'm pooped.

Last edited by kilaueabart on 22-Mar-2015 at 07:12 AM.

@kilaueabart

The CF card or USB flash drive must be 'ejected' after copying the kernel.

_________________
www.amigalinux.org

@kilaueabart

Er.... what xeno said +

Quote:

Loader:elf Filesystem:raw Dev:cf0 File:vmlinux-3.18 Options:(null) Loading: Failed. Could not load cf0:vmlinux-3.18: Not an ELF-format executable *** command status = -11 As I was typing those lines I vaguely remembered having intermittent trouble with kernel 3.18 before, and I think I may have time to give 3.14 a try tonight, as long as I don't make too many typos redoing MENU_3_COMMAND.

The only thing I can think of right now is that maybe(?) you forgot to unzip the kernel(?).

Instead of retyping the Menu item each time just enter the commands separately like so;

setenv bootargs 'root=/dev/sdb1' (Enter)

boot -elf -noints -fatfs cf0:vmlinux-3.xx (Enter)

And if you make a typo on the second line just hit the UP CURSOR key, backspace, and try again. After you figure out a proper sequence you can set it up for the Menu.

_________________
Tinkering with computers.

@Signal

After removing a lot of junk as advised by Christian (aka Xeno74), my CF0: (aka KINGSTON:) looks like this:

.Trash-1000 Dir ---arwed 28-Oct-14 15:19:34
vmlinux-3.12 18338734 ----rwed 04-Nov-13 04:32:44
vmlinux-3.14 18584865 ----rwed 30-Mar-14 19:52:36
vmlinux-3.17 21571096 ----rwed 05-Oct-14 18:33:12
vmlinux-3.18 21747363 ----rwed 21-Oct-14 04:52:10
vmlinux-3.19 22101199 ----rwed 02-Feb-15 03:38:52
5 files - 97M bytes - 1 directory - 200213 blocks used

As I recall, I have simply dearchived things like
vmlinux-3.19.0-rc7-KVM-PR-AMIGA_one_X1000.tar.bz2 8070466 ----rwed 05-Feb-15 14:08:09
to RAM: and then copied what I thought I might need to KINGSTON:. I have no idea how I could eject that card, except by opening the computer and sliding it out. I doubt that is what Christian means.

@Signal

Quote:

No, MBR is pretty much standard. Amiga RDB is the oddball.

Actually for Amiga PPC Linux in the last ten years, MBR would be an oddball, as RDB has had kernel support for ages and Linux booted quite easily off an RDB volume without funny corruption

For a whole HD dedicated to Linux would be a different matter. But still MBR for PPC seems like a foreign concept.

@kilaueabart

Quote:

I was optimistic that it would run from CFE. The attempt took quite a while because I simply cannot type one of those "setenv -p MENU_#_COMMAND ..." things without one typo, a different one each time I redo it. But I finally got a command the machine was willing to try, until soon it told me:

Do you have an existing Linux menu you can boot from? If so I suggest modifying it to boot a generic kernel file. That is set it to load a file like vmlinux. Then copy the kernel you want to test, say vmlinux-3.18 as vmlinux. Then you only need to set up the menu kernel once and can update the kernel with ease.

The only thing is, I don't recall if there are some "cheats" on CFE. You can print the variables to screen. But sticking a cursor over it to edit it is another matter. IF you enter al long command but mess it up you can at least cursor up and fix it.

As to this ejecting, what is meant is to dismount the drive, to ensure a clean file-system. IOW to safely remove it, so not a physical removal, but a file-system removal so the entire disk has been closed like a file. Now AmigaOS doesn't implicitly dismount external drives like those on USB. And it is usually okay to wait until disk activity has finished. But I don't know if that is good enough in this case.

CFE seems to be sensitive to the particulars of the FAT file system and if it looks okay in the OS or even to CFE as a dir listing it could still break in CFE when it tries to read a file.

You could try using the USB commodity (Ctrl-Alt-M) to dismount the drive if it supports that, Should be possible.

Last edited by Hypex on 23-Mar-2015 at 12:09 PM.

@kilaueabart

Quote:

kilaueabart wrote: @Signal After removing a lot of junk as advised by Christian (aka Xeno74), my CF0: (aka KINGSTON:) looks like this: .Trash-1000 Dir ---arwed 28-Oct-14 15:19:34 vmlinux-3.12 18338734 ----rwed 04-Nov-13 04:32:44 vmlinux-3.14 18584865 ----rwed 30-Mar-14 19:52:36 vmlinux-3.17 21571096 ----rwed 05-Oct-14 18:33:12 vmlinux-3.18 21747363 ----rwed 21-Oct-14 04:52:10 vmlinux-3.19 22101199 ----rwed 02-Feb-15 03:38:52 5 files - 97M bytes - 1 directory - 200213 blocks used As I recall, I have simply dearchived things like vmlinux-3.19.0-rc7-KVM-PR-AMIGA_one_X1000.tar.bz2 8070466 ----rwed 05-Feb-15 14:08:09 to RAM: and then copied what I thought I might need to KINGSTON:. I have no idea how I could eject that card, except by opening the computer and sliding it out. I doubt that is what Christian means.

Like Hypex said he probably means unmount.

It looks like your command line is different than mine.
Quote:

Loader:elf Filesystem:raw Dev:cf0 File:vmlinux-3.18 Options:(null) Loading: Failed. Could not load cf0:vmlinux-3.18: Not an ELF-format executable *** command status = -11

The first line has 'Filesystem:raw' and mine uses 'Filesystem:fat'.

Perhaps that is where the problem booting is?

I believe your boot command should be,

boot -elf -noints -fatfs cf0:vmlinux-3.xx

Again I suggest you enter the commands from the CFE prompt directly until you get this train moving.

_________________
Tinkering with computers.

@kilaueabart

This is the right output:

Quote:

Loader:elf Filesystem:fat

_________________
www.amigalinux.org

@kilaueabart

Please try it, how Signal has described it. With:

Quote:

boot -elf -noints -fatfs cf0:vmlinux-3.xx

Good luck.

_________________
www.amigalinux.org

@xeno74
All right, I'm off to try it in a moment, although I don't see how doing the "boot ..." separately from "bootargs" could help.

Before I go, just in case any of this matters, let me report:

Information > Disk from the KINGSTON (cf0:) icon tells me, among other things,
" Handler: CrossDOSFileSystem 53.11 (FAT2)"

Media Toolbox tells me:
" CF CARD [long hex number] ver7 0,0 Fixed hard disk Not installed"
The buttons are all ghosted, except Details, which opens an empty window

Control^Alt^M brings up a USB Mass Storages window that doesn't list anything at all.

OK, I've been to CFE and back. Interesting.

I did the setenv bootargs and boot separately, twice, and got pages of messages before it finally stopped with a "rebooting in 180 seconds." I managed to write some of it down, but 180 seconds is LONG UNLESS you're trying to use it. That's why I had to do it twice. These may not be in exact order:

VFS: Cannot open root device "(null)" or unknown-block (0,0): error -6
Please append a correct "root=" boot option, here are some available partitions
[first, all 9 of sdba partitions are listed]
0810 ... sdb driver sd
[indented]0811 ... sdb1 d73a0500-01 (/dev/sdb1 was of course my root= option every tiime)
[indented]0812 1 sdb2
[indented]0815 ... sdb5 d73a0500-05
Kernal panic - not syncing: VFS Unable to mount root fs on unknown-block (0,0)

The first time through it stopped there, and the screen went black. X1000 reset button had no effect, but maybe Ctrl^LA^RA, among the keys I tried, revived it! The second run ended with Call Trace: and about eight lines of numbers and words. This time the reset button worked.

I supposed the reason I was getting all those messages might be because I left "quiet" out of the bootargs, so I tried it once with "quiet," but no "ro" or "splash." Brief messages too fast for me to remember, soon gone, with some kind of quick notice that it was loading Ubuntu! Then something about a low resolution screen followed by an alert:

What would you like to do?
--------------------------------------
[checked] Run in low graphics mode for just one session
[Unchecked] Reconfigure graphics
[Unchecked] Troubleshoot the error
[Unchecked] Exit to cosole login
------------------------------------------------
Cancel / OK

I couldn't do anything. "Enter" had no effect, and there was no mouse error to select with. So I gave up until one of you suggests what I should try next. (Maybe reformat cf0: and reload?)

Last edited by kilaueabart on 24-Mar-2015 at 01:58 AM.

@kilaueabart

Could you write down your boot commands (boot/setenv), please?

If I understand you correctly, you can boot the installed Ubuntu Live Remix distribution, don't you?

_________________
www.amigalinux.org

@kilaueabart

I can reproduce the same error messages you are getting;
Quote:

VFS: Cannot open root device "(null)" or unknown-block (0,0): error -6 Please append a correct "root=" boot option, here are some available partitions [first, all 9 of sdba partitions are listed] 0810 ... sdb driver sd [indented]0811 ... sdb1 d73a0500-01 (/dev/sdb1 was of course my root= option every tiime) [indented]0812 1 sdb2 [indented]0815 ... sdb5 d73a0500-05 Kernal panic - not syncing: VFS Unable to mount root fs on unknown-block (0,0)

Simply by forgetting to put "root=/dev/sdb1" in quotes.

setenv bootargs "root=/dev/sdb1"

_________________
Tinkering with computers.