- This topic has 38 replies, 7 voices, and was last updated Dec 11-11:33 pm by christophe.
-
Posts
-
olsztyn
OK, the guy who has posted his findings at stackexchange explains, you’ll have to clean the stick completely except for the ldlinux.sys file and rewrite all files and directories at once from a backup copy to make syslinux work again on the stick. S
Is the original boot folder where ldlinux.sys resides named ‘boot’? Just to make sure… The config file with the name either syslinux.cfg or extlinux.conf must reside in that folder. There are a few other files from syslinux package that must be there, such as *menu* files (I think two).
Would this be not much distraction for you to post content of the syslinux.cfg here?
I think you are getting very close to make this work with antiX 21 and if I understand it now initiates booting (although not complete yet), which had not done before with a typical stick produced with live-usb-maker, although I still do not understand why, because syslinux ldlinux.sys was still in the same partition, although linux kernel in a different partition. Now all is in the same partition, so perhaps this makes a difference. This is just my vague conjecture based on the fact that the older version of syslinux (not extlinux) had to be on Fat32. So aas antiX live-usb-maker produces composition, where linux kernel is on a separate ext4 partition, this might have been a factor…
All what I am musing about here is just my rambling full of vagueness as I am quite confused how exactly it works with separate partitions and therefore I performed a separate experiment (posted in a separate topic), where I put all antiX in the same ext4 partition, got rid of antiX UEFI partition and it boots including UEFI…Live antiX Boot Options (Previously posted by Xecure):
https://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parameters
christophe
OK, Robin. I see you aren’t letting this go, and neither is olsztyn. And I believe I have a (potential) solution, too.
You were wondering if syslinux on a FAT partition alone would work. But the file sizes are limited (linuxfs, rootfs, & homefs) if we use that, and permissions problems, etc.
So I figured out how (on a 32-bit computer here) how you can (hopefully) “trick” the computer to boot with syslinux on a FAT partition, then “chainload” to a “secret” ext4 partition with antiX 21 frugal.
I made a prototype on stick, and it works for me. I have it available for you to download it here (just dd it to a stick 2gb-or-larger usb stick – If you go larger, then you’ll be safe from not having it potentially “cut-off” at the end, since manufacturers are always so inaccurate about exact sizes). Note: I may not be able to keep it available for a long time. But I’ll keep it up for a week or so, at least…
https://www.dropbox.com/s/217x87jat22lf91/a21-syslinux-usb.img.tar.gz?dl=0
(I only hope it works for you on your finicky bios.)
Here are the instructions to recreate this yourself:
My train of thought was just that, since what you have works, and syslinux is booting it, then if I were to have the boot partition be the same.. only then, once it has booted, chainload into ext4 & linux.
Obviously, adjust device names, as applies to your setup.
1. In gparted, make the partitions
a. I made a new msdos partition table on the usb.
b. 50 mb fat16 partition as sda1, labeled BOOT
c. all the rest as sda2, ext4, labeled Frugal
d. after tasks completed, mark the (sda1 in my case) FAT partition flag as “boot”. Close gparted.EDIT: At this point, copy the “antiX” directory from the root directory on a live-CD/usb, to the ext4 partition. Rename it — mine is antiX21.
2. Create /boot/extlinux directories in sda2 (ext4), then unmount the stick’s FAT partition (but leave plugged in). Then run as root (su/sudo):
a.syslinux --install /dev/sda1
b.dd if=/usr/lib/SYSLINUX/mbr.bin of=/dev/sda bs=440 count=1
c.extlinux --install /media/demo/frugal/boot/extlinux3. Copy over all the files (as root) from /usr/lib/syslinux/modules/bios to BOTH the root of the FAT partition (mounted now as /media/demo/BOOT in my example), and the antiX ext4 partition (mounted as /media/demo/frugal/boot/extlinux for my example).
4. Making the syslinux.cfg on the FAT partition. Copy this to the root of the FAT partition, noting that the chainloading command is referencing the 2nd partition of the usb.
UI menu.c32 PROMPT 0 MENU TITLE Boot Menu TIMEOUT 50 DEFAULT default LABEL default COM32 chain.c32 APPEND hd0 25. The /boot/extlinux/extlinux.conf file needs to be on the ext4 partition. It’s straight-forward from the frugal-extlinux thread referenced earlier.
UI vesamenu.c32 TIMEOUT 100 MENU HIDDEN MENU TITLE antiX Linux frugal MENU BACKGROUND splash302.png MENU COLOR border 30;44 #40ffffff #a0000000 std LABEL l0 MENU LABEL antiX 21, static persistence LINUX /antiX21/vmlinuz root=UUID=a50eab49-829b-4087-968b-12289ca9b944 bdir=antiX21 buuid=a50eab49-829b-4087-968b-12289ca9b944 APPEND p_static_root vga=default quiet disable=lF INITRD /antiX21/initrd.gz6. You can add the splash image (splash302.png) to your /boot/extlinux directory.
(I think it looks nice.) Press any key to see options & stop the automatic boot in 10 seconds to antiX-21…Anyway. I hope this works for you. 🙂
- This reply was modified 1 year, 5 months ago by christophe. Reason: added code blocks
- This reply was modified 1 year, 5 months ago by christophe. Reason: clarification
- This reply was modified 1 year, 5 months ago by christophe. Reason: Forgot a step - copying the frugal directory
Attachments:
confirmed antiX frugaler, since 2019
Robin
@christophe I can’t let it go if I want to see antiX 21 running.
Many thanks for your image, but it won’t boot here. Same problem as with the one I created myself following your instructions. It simply isn’t accepted as a valid boot medium. I’m working on changing this also (I have a faint idea already what could help out), but this is another tree to climb. First I’ll follow the path using the dd’ed drive image and replacing the files and folders on it to run antiX 21 instead of antiX 17. At least this 1:1 copy of the old stick boots fine and is accepted by BIOS as bootable medium.
I investigated a bit further and found probably the reason for all the trouble: The old style BIOSes only accept MS-DOS media as bootable from USB RMD-FDD. (My USB sticks are not accepted in BIOS boot mode “USB-hdd” by now) Remember, I had to activate the feature “USB support for DOS” to make it boot at all, so it expects true ms-dos style filenames, menaning I’ll have to control exactly the writing of the correct 8.3 filenames. And unfortunately exactly this is something the linux vfat driver is not capable of:
An answer to this issue I found on another board:
»Although the FAT driver in the Linux kernel provides an ioctl to list the short filenames in a directory, it provides no ioctl to set the short name of a file. This means that lossless backup and restore of short names is not possible when working through the FAT driver in the Linux kernel.«
This is at least the answer to the strange haunting and vanishing of files and directories when moved and renamed using antiX (or any linux) on a vfat file system. The 8.3 filename won’t get written at all to disk, so clear chain of reasoning is syslinux can’t find them on boot. So I’d state the linux vfat kernel driver is considered as broken actually. I’d only like to come to know how the guys from unetbootin have managed this once when I created my very first antiX stick on ubuntu… Maybe they write these names directly to the disk, bypassing the broken linux vfat driver. But this is pure guesswork.
So also the rewriting of the complete content all at once to the blank USB stick, containing only ldlinux.sys, as suggested by the guy on stackoverflow, didn’t work. Understandable, the 8.3 filenames are still missing this way…
So maybe I could manage to modify the file content of the files in the existing directory without renaming or recreating them, so they will be still seen on boot time by syslinux, keeping the original 8.3 names in place. Possibly I’ll have to boot from an ancient 5¼” MS-DOS diskette in order to inspect which true 8.3 filenames exactly are used in the original bootable USB stick. I hope this old drive will work still, I’ve never used it for years. Hadn’t imagined I’d ever use it again. But on the other hand I wouldn’t have expected to see the linux vfat driver still broken, forcing me now to read the names from within a true MS-DOS environment. This is something I was used to in the very first days of linux, but… We are decades later meanwhile! The vfat driver should really have got fixed in our days. Or does anybody know how to see the true 8.3 vfat filenames as written on disk from within linux? The quote above reads: “…the FAT driver in the Linux kernel provides an ioctl to list the short filenames in a directory,…”. But how is this done? It seems I’m not able of seeing these names from within antiX. Even mc doesn’t show them correctly.
Windows is like a submarine. Open a window and serious problems will start.
@olsztyn per your request:
Original 17.4.1, boots fine:
#-------------------------------------------------------------------- # This is the isolinux.cfg and/or syslinux.cfg file # It controls the main menu in the bootloader on the live system. # You can edit it to change the main bootloader menu on a LiveUSB. # If you are not careful you can break the live system and prevent # it from booting. #-------------------------------------------------------------------- UI gfxboot gfx-cpio readme.msg timeout 3000 default live MENU TITLE Welcome to antiX-17.4.1_386-full (Helen Keller) LABEL live MENU LABEL antiX-17.4.1_386-full (28 March 2019) KERNEL /antiX/vmlinuz APPEND quiet splash=v disable=lx INITRD /antiX/initrd.gz LABEL safe MENU LABEL Safe_Video_Mode KERNEL /antiX/vmlinuz APPEND quiet splash=v disable=lx nomodeset xorg=safe INITRD /antiX/initrd.gz LABEL failsafe MENU LABEL Failsafe_Boot KERNEL /antiX/vmlinuz APPEND quiet splash=v disable=lx nomodeset failsafe INITRD /antiX/initrd.gz LABEL harddisk MENU LABEL Boot_from_Hard_Disk COM32 chain.c32 APPEND hd1 LABEL memtest MENU LABEL Memory_Test KERNEL /boot/memtestReplacement from antiX 21, refuses to work. But as explained in the posting before, this is caused by missing 8.3 filenames when writing these files from within antiX to the fat32 boot medium, not by this file or its content itslf. It also fails when creating a fresh 1:1 copy of the original 17.4.1 syslinux.cfg file as shown above using the same correct file name in the very folder.
#-------------------------------------------------------------------- # This is the isolinux.cfg and/or syslinux.cfg file # It controls the main menu in the bootloader on the live system. # You can edit it to change the main bootloader menu on a LiveUSB. # If you are not careful you can break the live system and prevent # it from booting. #-------------------------------------------------------------------- UI gfxboot gfx-cpio readme.msg timeout 3000 default MENU TITLE Welcome to antiX-21_386-full (Grup Yorum) LABEL live MENU LABEL antiX-21_386-full (31 October 2021) KERNEL /antiX/vmlinuz APPEND quiet splasht disable=lxF INITRD /antiX/initrd.gz LABEL safe MENU LABEL Safe_Video_Mode KERNEL /antiX/vmlinuz APPEND quiet splasht disable=lxF xorg=safe INITRD /antiX/initrd.gz LABEL failsafe MENU LABEL Failsafe_Boot KERNEL /antiX/vmlinuz APPEND quiet splasht disable=lxF failsafe INITRD /antiX/initrd.gz LABEL harddisk MENU LABEL Boot_from_Hard_Disk COM32 chain.c32 APPEND hd1 LABEL memtest MENU LABEL Memory_Test KERNEL /boot/memtest LABEL grub MENU LABEL Switch_to_Grub_Bootloader KERNEL /boot/grub/i386-pc/lnxboot.img INITRD /boot/grub/i386-pc/core.imgas I am quite confused how exactly it works with separate partitions
From reading this text you’ll probably get an idea how all this works in detail. All syslinux boot stages are described, and how they depend on each other. Also what which file is good for.
Windows is like a submarine. Open a window and serious problems will start.
This is at least the answer to the strange haunting and vanishing of files and directories when moved and renamed using antiX (or any linux) on a vfat file system. The 8.3 filename won’t get written at all to disk, so clear chain of reasoning is syslinux can’t find them on boot. So I’d state the linux vfat kernel driver is considered as broken actually. I’d only like to come to know how the guys from unetbootin have managed this once when I created my very first antiX stick on ubuntu…
I must say this is very interesting, unexpected scenario. I would have never figured this out… So congrats on discovering this…
Very curious how this was correctly created via Unetbootin on Ubuntu. And if Live-USB-Maker created the boot partition (49M) as Fat16 instead of Vfat, whether files would be written out correctly. Or if instead of 64Gb Fat32 partition to compose such antiX21 stick the first partition to contain antiX boot files is set up as Fat16 (4Gb) instead of Vfat. This means whether this would make a difference in behavior of driver.However you will end up resolving this puzzle, I trust you will find an easier solution than using old MS DOS…
This scenario however brings up an idea to build-in more options into Live-USB-Maker and ISO Snaphot, to be able to specify a file systems and partitioning scheme when creating antiX Live, such as whether to use the Vfat 49Gb ANTIX_UEFI partition or make it a one-partition system, and what file system to use…
I hope antiX team and owner reads these posts and for me personally your experience and discovery of this phenomenon is a great lesson to learn…- This reply was modified 1 year, 5 months ago by olsztyn.
Live antiX Boot Options (Previously posted by Xecure):
https://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parametersseaken64
Your antiX-21 config file is missing the “Default” label. Try “default live”.
Have you tried Plop! boot manager? It can be started from floppy in DOS mode. There is a settings file that can be used to change USB modes.
Just an idea.
Seaken64
Personally, I would not attempt to install Linux at all on a critical work machine that I cannot afford to lose. It would be better to either remove the hard drive and install a different hard drive, or use another machine.
I know you have your reasons. I am just commenting in general as to the risk of losing your work machine by playing with Linux on that same machine. I run a business and I have installed Linux thousands of times. Things go wrong and the business gets affected if we make a mistake. Just my two cents worth.
Seaken64
christophe
Where are you on this, Robin?
I’m thinking that USB ARMD-FDD makes a lot of sense for the era — atapi removable media device. Some of those were 100-120 MB at the time…
I’m looking at some alternative boot methods. But I don’t want to duplicate something you’ve considered settled as not viable… 🙂
confirmed antiX frugaler, since 2019
As a similar case for me to study and pursue booting antiX USB on two of my old desktops I retrieved from storage:
– Compaq Evo – Pentium 4 based small format horizontal desktop with 10Gb IDE hard disk
– Self made from parts tower game machine with AMD Phemium, multiple disks, 8Gb ram, powerful Radeon graphics card
Experiment:
– On the first machine I have not found so far a way to boot antiX Live, but will keep playing with settings. So far I ended up taking the HD out, connecting to the other (tower) machine, booting antiX Live on that one (see below on this) and installing Live instance to HD (–force=usb parameter on live-usb-maker). Such antiX Live is back in Compaq Evo and booting antiX Live from HD.
– On the tower at first I could not make it boot from antiX Live usb stick. Boot continued to go to HD in spite of removable boot priority in BIOS. Puzzled I displayed disk storage configuration and noticed my usb stick (PNY) was listed as hard drive (although it is a regular removable, with no fixed HD bit set) being second in boot priority, so kept booting from first (real HD) and not usb stick emulated as HD. Switched booting priority to PNY usb stick (showing as HD) and booted antiX Live just fine. So the issue booting from usb stick was that boot was emulated as HD (second in boot priority though) not removable usb media, so it boot priority adjustment was among HD devices for it to boot first.
Having booted antiX Live stick on the tower machine I installed antiX Live on the HD from Compaq (command line Live-USB-Maker –force=usb) and put back in Compaq machine.This case probably does not apply to Robin’s case, as that usb stick is not being emulated as HD under any circumstances as I understand…
My next challenge will be making antiX Live USB stick to boot on the Compaq Evo machine (currently it runs antiX Live from HD)…- This reply was modified 1 year, 5 months ago by olsztyn.
- This reply was modified 1 year, 5 months ago by olsztyn.
- This reply was modified 1 year, 5 months ago by olsztyn.
Live antiX Boot Options (Previously posted by Xecure):
https://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parameterschristophe
@olsztyn – good idea on moving the hard drive for the P4. Probably can’t boot from usb at all on that one, or is it similar to Robin’s? (I have an “Athlon XP 2200+” from 2002. CMOS battery is dead – the bios resets the date to 01/01/2002 whenever I unplug it. It can’t boot from usb at all, unfortunately.)
confirmed antiX frugaler, since 2019
Hello all, I’ll be back on this task this evening. Sorry for the delay in answering, first I had to complete the writing of script code for automatic translation of anitX desktop files to the languages we cover, which have no real translators by now, and also I had to check for the unplugdrive woes of antiX 21 version which was heavily modified last minute from 0.91 to 0.93. Xecure did a great job in acceleration of processing performed by unplugdrive, but there was no time left for testing it as extensively as I did while I have written the 0.91 version, on all kinds of devices before it went into the ISO. So we face now the absurd situation we have a well tested 0.91 (slow and sudo) version in the archives, and a mostly untested 0.93 (fast sudo+nonsudo) version in antiX 21 🙂
But back to my antiX Live USB struggles.May be I still not reached at the bottom of reasons for syslinux not seeing the existing config files at boot time. I copied the files from within true Win2000 to the stick, which should write the correct 8.3 filenames. But still syslinux won’t see the config files.
My next approach was to copy the complete content of ../boot/syslinux subfolder to the root of the USB stick, even if they write in their manual syslinux would check the ../boot/syslinux and ../syslinux folder BEFORE checking the root directory for the config. Surprise: Now syslinux comes up all of a sudden with the antiX 21 boot screen. What the heck? I really don’t see the logical reasoning behind this, since the original config of antiX 17.4.1 was also located in the very subfolder, not in the root. So first (partly) success: I can boot the stick on this PC now and I saw the antiX 21 boot screen the first time. The localised F1 help looks great, @anticapitalista you did a great job with allowing the help menu to be localised now. This makes antiX usable to many more people all over the world. Many thanks for what you do for all of us!
OK, but now back to the LiveUSB. Again the strange “file/folder not found” message hit me, this time /antiX/vmlinuz file is not found while it is actually present in this folder. Slowly but surely I’m fed up to the back teeth by this stupid message. The files ARE present in the directories.
Btw: Doesn’t mean syslinux “sees” the files when they sit in USB root, that all the filenames fine, and merely all new folders are “invisible”?
Just an idea… Since this is a MS-DOS style file system, does syslinux possibly expect the slashes to be written the other way around in its config file now for some strange reason? I mean \ instead of / ? But why does all this work flawlessly in the original 17.4.1 stick then, trouble starting the minute a folder is created new? It must be somehow interdepend on the way how the folder and file names are written to the file system. I have not tried to do this on true old school MS-DOS instead of antiX or Win2000, maybe this would make a difference. Once we have understood WHAT exactly prevents syslinux from seeing the folders at boot time, we can try to treat it while creating the stick from within antiX…
@seaken64 Many thanks for your warnings. But I’m well secured, there is a 1:1 backup hdd, function tested and ready to plug, and in case something unexpected happens to the CAD station, it is a task of 10 minutes max only to make it run again exactly as before. Just switch the cables within the PC case from the hdd in the one slot to the spare hdd in the other slot. Moreover all local user data is stored on external USB drives, internal hdd containing the pure OS and CAD system only. I won’t install antiX on this PC, it is to be run from persistent Live USB stick exclusively, and antiX is meant for allowing external data exchange on this machine, while I wouldn’t never ever let it connect to network or internet when running it on its outdated Windows, Microsoft not providing any security updates anymore and not being able to offer a functioning replacement version of their OS running on this hardware.
Windows is like a submarine. Open a window and serious problems will start.
good idea on moving the hard drive for the P4. Probably can’t boot from usb at all on that one, or is it similar to Robin’s? (I have an “Athlon XP 2200+” from 2002
This Compaq Evo D500 is a challenging machine, great for testing antiX… From inxi: Pentium III Celeron Tualitin, 1.3 Ghz, 256M ram…
Manufactured in 2001 it seems and USB is 1.1…
Running now antiX 21 Runit… Not bad being 2001 low end machine…Live antiX Boot Options (Previously posted by Xecure):
https://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parametersstevesr0
Three suggestions:
(1) There is a related thread (https://stackoverflow.com/questions/15624772/debugging-the-boot-filesystem-environment-seen-by-syslinux) that might be helpful. I am not technical competent to pick out the good parts. It is possible you folks will find a hint to a solution there.
Two possible work arounds (I realize even if this works, it doesn’t fix the underlying problem.):
(2) there are programs (plop bootloader is one, website is plop.at) that enable a usb flash drive to boot on a machine that couldn’t boot from usb flash drive.
(3) I see a post from 2013 in the PCLinux OS forum that said if you go to the boot menu in the BIOS and select hard drive with the usb flash drive attached, you will get a choice of the hard drive and the usb stick. If you move the usb drive above the hard drive then the usb drive will boot. This is different than launching the boot selection menu – which is said not to work. Subsequent commentors said this doesn’t always work.
stevesr0
- This reply was modified 1 year, 5 months ago by stevesr0.
I see a post from 2013 in the PCLinux OS forum that said if you go to the boot menu in the BIOS and select hard drive with the usb flash drive attached, you will get a choice of the hard drive and the usb stick. If you move the usb drive above the hard drive then the usb drive will boot.
Yes. This is exactly what I experienced on my desktop system, booting I described in my post #72437 above. Desktop BIOS sees USB stick as an HD, below existing HD in boot sequence. So I move the usb stick ‘HD’ up in boot priority above existing disk and it booted happily after…
My experiment with desktops I brought back to experimental life was mostly to recall how to make them boot from USB removable drive, where such boot option might not be apparent in their BIOS and now I think most BIOSes had HD emulation, more common.
I cannot help thinking that in Robin’s case there must be a simpler way to boot. So first would be to double check if USB stick is not being listed among HD drives…- This reply was modified 1 year, 5 months ago by olsztyn.
Live antiX Boot Options (Previously posted by Xecure):
https://antixlinuxfan.miraheze.org/wiki/Table_of_antiX_Boot_Parameters@christophe
I have an “Athlon XP 2200+” from 2002
This is nearly the very specs of this PC here, AMI BIOS 12/20/2002 K7S5A PRO release, 2GB RAM, Athlon XP 2000+
@stevesr0
Probably this thread is quite long meanwhile, the thread you’ve linked to on stackexchange is exactly the one I referred to, from which I drew most of the infos already.(2) and (3): The problem here is not the PC wouldn’t boot from a USB device. It boots fine using the old antiX 17.4.1 USB stick. Even when DDing this stick 1:1 to a fresh one, this copy will boot fine also. So there should be no need of obscure techniques to betray the BIOS to make it boot from USB. The Problem I face here is: There obviously has changed something in the way the USB sticks are written by linux OS nowadays, so they won’t be accepted on the old machines anymore. I recall it was a task of no time to create the 17.4.1 stick formerly, and it ran out of the box immediately.
Last findings:
There is an installer flag accepted by syslinux, which makes any of my Test sticks recognised immediately by this PC when updating the syslinux installation on the flash drive, as long as it is FAT32 formated and has a single partition.
Add the argument »–stupid« to the line and the stick will never fail to come up with the messagessearching for Boot Record from USB ARMD:FDD… OK
..
SYSLINUX 6.04 CHS 20190226 copyright (c) 1994 – 2015 Peter Anvin at alIt also “sees” the config file now, since there are no complaints anymore about “Warning: No configuration file found”, at least as long I put it alltogether into the root directory of the USB stick. But now it comes up with new ideas:
No DEFAULT or UI configuration directive found
Boot:I double checked: The lines are present within the file.
Why does all this work fine on the 17.4.1 USB stick, and the moment I replace whatever in the configs or use the file set from antiX 21 instead on the very stick it fails at each next step again and again? What have they done to this bootloader, that it stops working now? Looks like some kind of acceleration activated the old style PC can’t handle?
Where are you on this, Robin?… But I don’t want to duplicate something you’ve considered settled as not viable…
What I did until now was:
– Copy the complete content to the USB stick at once, as described by the guy from stackoverflow.
– Check each of the different USB boot modes this BIOS provides. The single one coming up with the syslinux boot loader message at all is “ARMD-FDD” mode. When using the old 17.4.1 Stick it boots fine this way. Replacing the files or modifying the configuration in a 1:1 copy of this stick makes it fail. Only once I managed to see the antiX 21 boot screen, but after trying to correct the paths within the config file to make it load the kernel actually when having performed all selections this screen provides it doesn’t show up anymore, not even when replacing it by a fresh copy of the original config file. I’d have again to dd from the original stick first to make it come up again.
– Tried to copy hdt.c32 and rosh.c32 modules from /usr/lib/syslinux/modules to the boot medium and use it from boot: comandline. “File or directory not found”.
– I compared the mbrs of the old and the new stick within hex editor. They differ in 2 bytes only, but copying the old mbr to the 21 Stick doesn’t change things. Checked Byte 0x01FE and 0x01FF to be set to 55 AA. Can’t find any error.
– A real advance is using the “–stupid” argument when installing syslinux. I also applied this to the original Stick created by antiX Live USB maker for antiX 21 (extlinux does know this also), but the stick designed by antiX isn’t accepted as a valid boot medium nontheless. Probably the BIOS can’t handle multiple partitions or ext file systems for boot. I’ll have to create a stick containing an FAT32 DOS partition instead of an UEFI Partition, likewise chaining to the big ext4 partition for further boot after syslinux is running already. Possibly this could help out.
– Changed the direction of slashes within the config files, in case it is expected to be MS DOS style for some reason all of a sudden.
– Moved the all content to root directory and corrected the paths within config file.
– I’ve used the syslinux tool mkdiskimage to create an USB stick in HDD style, to make it accepted by BIOS as a real USB HDD. Starting at Block 63 instead of 2048 and showing 255 heads and 63 sectors on 7648 Cylinders, LBA mode and Boot flag set, formatted FAT32. (BIOS recognises still ARMD-FDD.) Syslinux comes up this way still, but doesn’t find its file ldlinux.c32 anymore now and complains, while it sits in the very folder.
– Finally I checked the file system of the USB sticks on a true Windows 2000 machine, using its System management tools. Everything fine, no errors at all. Even defragmented the drive, in case fragmentation of files or folders would prevent syslinux from seeing them on boot time.
– Next attempt was to rewrite all the files from within Win2000 to the stick instead of using the linux vfat driver. But not even this made antiX 21 live USB boot.
– Finally I installed unetbootin on the antiX 19 machine. All in vein, the recent version as of antiX 19 doesn’t create the stick in a way that it is bootable on this machine, as the sticks I have created long ago using the same tool on ubuntu still. I can’t understand why this won’t work anymore.As said more than once, the antiX 17.4.1 persistent USB stick still works fine (including saving of root and home persistence and copying its stored content to ram on next boot) and can even get duplicated by dd (not by antiX tool) to another fresh USB stick, which will also boot fine then. It’s the very stick which refuses to boot when the files are replaced on it, or when it is written completely by antiX. So the stick itself is OK and the PC is proven to be able to boot from it. It really should run with the same setup on antiX 21 also…
So this is the most recent state of affairs here.
Windows is like a submarine. Open a window and serious problems will start.
- You must be logged in to reply to this topic.
