Multi-Porteus-stick Openbox boot problem [SOLVED]

[Amaya Nidjit]

Hello Again All.
Another request for help!

I made a Multi-Porteus boot stick with 4 versions on it, (supposed to make it easier to work on the other problem posted elsewhere).
Thanks much to the writer of that How-To explanation http://www.porteus.org/tutorials/10-ins ... ition.html..

Porteus versions with XFCE, LXQT, and MATE desktops all boot fine from the stick, but the Openbox version refuses to boot, claiming that the "from=" cheat code is invalid. However, that parameter is in the same FORM as for the boot instructions for the other Porteus versions, all of which boot OK.
(relevant portion of /Popenbox/boot/syslinux/porteus.cfg below)

#modified from provided code to suit Amys setup
LABEL openbox # The label 'openbox' is arbtrarily provided - Im not really sure what this param is for.
MENU LABEL OPENBOX 32-bit desktop
KERNEL /Popenbox/boot/syslinux/vmlinuz
APPEND initrd=/Popenbox/boot/syslinux/initrd.xz from=/Popenbox changes=/Popenbox/porteus
# FROM parameter is same form as other successful boot entries, but wont boot: sez "invalid cheat code".
# All others Porteus versions work OK with same form param.
TEXT HELP
Run 32-bit Porteus with Openbox desktop from subdir /Popenbox
ENDTEXT

I checked spelling. I checked that there were no invisible leading or trailing spaces in the name. I tried including the full path. No Joy. I can't see what I am doing wrong here. Can anybody else see it?

Thanx, Amaya

[Ed_P]

This is a menu from my porteus.cfg file.

Code: Select all

LABEL GRAPHICAL
MENU LABEL Graphics mode
KERNEL /boot/syslinux/vmlinuz
INITRD /boot/syslinux/initrd.xz
APPEND changes=/porteus
TEXT HELP

See a difference?

[rych]

Wow, I didn't know 32-bit versions still existed Also, that How-To is VERY old. Anyway, it won't hurt to try different syntaxis of the "from=" cheatcode. How many partitions do you have on your disk? Surely your have a FAT32 for UEFI booting, and the main one, is it ext4? If so, take a look at my multi-boot setup (described in [HOWTO] Multiboot Windows and frugal Porteus/X)

menuentry 'Porteus' {
linux /vmlinuz changes=/porteus login=root from=LABEL:PorteuSX fsck zram=33%
initrd /initrd.zst
}
menuentry 'PorteuX ' {
linux /vmlinuzX from=LABEL:PorteuSX changes=/porteux login=root fsck
initrd /initrdX.zst
}

I just have /porteus and /porteux folders on an ext4 partition *labelled* PorteuSX. I think modern versions of Porteus and PorteuX can then find those folders automagically even without a from= cheatcode!? But in the past at least, explicitly helping it with e.g. from=LABEL:<partition-label> was necessary...

[rych]

Another thing to check would be the integrity of the /Popenbox/porteus folder. Does it actually have all the right subfolders, including an (empty) /Popenbox/porteus/changes?

[Amaya Nidjit]

Thanks rych & Ed_P for your responses.

@Ed_P: Yes, your post is the standard default cfg format as the installer writes it. But I can't use that exactly, because my flavours are in folders each their own. I don't see any SIGNIFICANT difference - other than those that are required. The from= parameter appears to be required as soon as /porteus is not in the root. This is almost always required (in various distros running multiboot frugally), that the starting location is specified somehow. I also tried the other (booting) Porteux (plural of Porteus) without it, and it unboots them.
Your sample code also specifies initrd in the more normal form I see. But passing the initrd location as an append parameter works the same, does not it? I see this not as an important difference - when I try it both ways in the other (working) Porteux boot codes, they boot either way.
The paths are of course different because each frugality has its own folder.

@rych: Yuppa, 32-bit Porteus is still offered for download. I find 32-bit stuff usually runs faster on 10 or 15-yr old 64-bit computers.

I tried a lot of different variations of the cheat code. and indeed, none of them hurt in the least :-). But alas, none of them helped either :-( The only one I have not tried yet is by UUID, because (I'm thinking) that just specifies the drive, and I've been assuming its the directory that is being not found. I will try that next.

The USB drive has a ~30Gb ext4 partion containing 4 frugal Porteux, each in its own folder and a ~100Mb fat32 partition for saving notes to share between the 4 flavours of Porteus, so I can compare the way the different flavours work, without having to rely too much on memory (mine, not RAM). There is no setup for UEFI booting. Its not needed here - none of our computers have UEFI enabled, and I am certainly not yet ready to try running Porteus on anybody else's computer. Its still too unknown to me.
I will look at your How-To. Thanks for the link.

You wrote...

I just have /porteus and /porteux folders on an ext4 partition *labelled* PorteuSX. I think modern versions of Porteus and PorteuX can then find those folders automagically even without a from= cheatcode!? But in the past at least, explicitly helping it with e.g. from=LABEL:<partition-label> was necessary..

OK, I learned somethng useful just from this paragraph. I didn't know how that LABEL function was working, and this pretty much explains the gist. I will be surprised, but pleased, if it turns out that finding the partition has been the problem. This, and the UUID spec should both have same effect?

Another thing to check would be the integrity of the /Popenbox/porteus folder. Does it actually have all the right subfolders, including an (empty) /Popenbox/porteus/changes?

OK, good point about malcopy.
Well the /Popenbox/porteus/changes folder is not empty, because I copied it wholesale from the preconfigured solo-boot drive I had been running for the last couple of weeks. Why lose all my customizations!? But I did the other three Porteux the same way, and they all boot up successfully with all the previous changes and files intact and functional. I copied the openbox flavour exactly the same way, and don't remember seeing any error messages that would indicate malcopy. However, I will follow your line of query, and try recopying all the original to the new stick - maybe it will work (and it won't hurt a bit) :-)

Thanks, Amaya

[Amaya Nidjit]

@rych...
Just read through your How-to. I can see that this would be very helpful for someone new to Linux and still hanging on to Windows for dear life.
But nobody in this house uses Windows!
I haven't used it "in earnest" myself since Vista EOLed.
(Off-topic, I don't know why people make such a big deal about switching to Linux - I just downloaded some ISOs, burned a platter for each, booted the platters LIVE and installed several on an old spare laptop, and started using. By the time Vista EOLed a year later, I was all comfy with Linux, knew my favourite distro of the bunch, and replaced Vista with it on the main box. So easy.)

Most of the multiboot thumb drives I make can be set up really easily with Grub4DOS (another reason I like 32-bits distros - G4D sometimes can't talk with 64-bits). When that doesn't work I use Grub2 or Syslinux, neither of which I am very comfortable with - as you can see from my difficulties with Porteus/Syslinux.
Ventoy also works like a hot damn for some distros, without knowing a thing about what you're doing. But for Porteus, the novel just-copy-and-boot installation method was new and fascinating, so I am sticking with that for my Porteus test drive.

Thanks, Amaya

[Ed_P]

I don't see any SIGNIFICANT difference

#modified from provided code to suit Amys setup
LABEL openbox # The label 'openbox' is arbtrarily provided - Im not really sure what this param is for.
MENU LABEL OPENBOX 32-bit desktop
KERNEL /Popenbox/boot/syslinux/vmlinuz
APPEND initrd=/Popenbox/boot/syslinux/initrd.xz from=/Popenbox changes=/Popenbox/porteus

INITRD /boot/syslinux/initrd.xz
APPEND changes=/porteus

You don't see INITRD on a separate line from the APPEND line?!

For your menu I suggest trying:

Code: Select all

INITRD /Popenbox/boot/syslinux/initrd.xz
APPEND from=/Popenbox changes=/Popenbox/porteus

Added in 17 hours 8 minutes 56 seconds:
instead of

Code: Select all

APPEND initrd=/Popenbox/boot/syslinux/initrd.xz from=/Popenbox changes=/Popenbox/porteus

APPEND is for adding cheatcodes and initrd= isn't one.

[Amaya Nidjit]

@Ed_P...
I think I accidentally hid the previous response to this in a TLDR. Apologies.
re:
INITRD /Popenbox/boot/syslinux/initrd.xz
APPEND from=/Popenbox changes=/Popenbox/porteus
instead of
APPEND initrd=/Popenbox/boot/syslinux/initrd.xz from=/Popenbox changes=/Popenbox/porteus

In brief, both work.
The append=initrd etc variant was prescribed by the multiboot how-to, so I used it.
I've never seen it before while editing boot files for Grub4DOS, but then, I ain't seen much.

Also, note that when you TAB to edit the Porteus command-line at syslinux boot screen, even the more usual INITRD separate-line variant gets displayed as if an APPEND (including any #comments on the same line, which is annoying). All very confusing. So I had to try them both (on the other flavours) and they both work.

Have not found time to check yet, but I am starting to think rych's suggestion of a malcopy, tho rare, would explain the no-boot issues. Blush.

Amaya

[Ed_P]

In brief, both work.

I'm truly surprised!

The append=initrd etc variant was prescribed by the multiboot how-to, so I used it.

Can you please provide a link to that how-to?

I did find an old forum posting which mentions initrd in the append line:

append initrd=/boot/initrd.xz quiet first_run reboot=efi,pci,force

but these instructions are for the old Kiosk version of Porteus, which is similar to the normal Porteus but is not based on slackware and is different.

[Amaya Nidjit]

Can you please provide a link to that how-to?

Sure. Here is the link to the DIY I followed ...
http://www.porteus.org/tutorials/10-ins ... ition.html

OK, now for the bad news...
This was much ado about the wrong thing - the cause of the boot failure was NOT anything to do with cheat codes after all, EVEN THO the syslinux error message SAID it was an invalid cheat code.
The real reason was a malcopy event, as suggested by rych.
When I checked, the filenames appeared in my file manager - so it looked like they were there. But if I clicked them, I got a snarky note that said "The File Does Not Exist".

So when I re-copied all the contents of the solo-boot Porteus stick (source) into the multi-boot stick's /Popenbox folder, it all worked fine on re-boot.

The post editor is malfunctioning obstructively, so I am going to continue this in another post.

[Amaya Nidjit]

OK, the editor seems to work fine in a new instance.

To finish up, I'd like to thank Ed_P and rych for the time they spent on this. I am embarrassed that I got it so wrong - but in my defense, whatever (syslinux or vmlinuz?) issued the erroneous error message that I believed, also got it wrong.

I may be grasping at strawberries, but to try to put a positive spin on this goosechase, something worthwhile came of it...
(1) A potentially useful-to-others How-To-Torial was brought to the fore.
(2) Ed_P, I, and (maybe) any others watching learned that the boot command sequences are more flexible and versatile than previously realized.
(3) I learned something the LABEL: function could do.
(4) rych learned that yes, there are still 32-bit versions of Porteus being released.
(5) and onlookers may also have been reminded that if you respond respectfully and helpfully even to what seems like a stupid or elementary question, everybody wins.

So I have marked this thread as SOLVED.

Amaya

[Ed_P]