Tiny Core Linux
dCore Import Debian Packages to Mountable SCE extensions => dCore X86 => Topic started by: theYinYeti on September 28, 2013, 08:25:34 AM
-
Hi! Would someone volunteer to post some kind of “getting started” instructions for dCore at this stage?
So far, I did this:
$ import xorg-all
[…]
$ printf 'Xprogs\nflwm_topside\nwbar\n' >stddesk
$ import -s -f stddesk
[…]
$ printf 'xorg-all.sce\nstddesk.sce\n' >>/etc/sysconfig/tcedir/sceboot.lstBut then, on boot, I am greeted by an endless output of this same line:
can't run '/sbin/getty': No such file or directory
I just don’t know how to continue from there, or what I did wrong.
Also, on each boot, I have this error message:
swapon: /dev/zram0: Invalid argument
Help! :D
My boot options are:
waitusb=5 tce=LABEL=FLASH/.boot/dcore/sce restore=LABEL=FLASH/.boot/dcore kmap=azerty/fr-latin1 loglevel=3 although the kmap is probably useless for now.
-
See: http://tinycorelinux.net/5.x/x86/README
In particular:
http://tinycorelinux.net/5.x/x86/README/README-1st.txt
http://tinycorelinux.net/5.x/x86/README/X-Desktop.txt
-
http://tinycorelinux.net/5.x/x86/README/README-X-Desktop.txt
(:
-
My thanks to both of you :)
I thought these files were related to regular TinyCore, with which I am mostly familiar already. I'll read these!
-
You were right, that was exactly what I was looking for.
Unfortunately, the instructions did not work for me. I did everything as explained, except that I set “tce=…” and “restore=…” beforehand in my Grub2 config file, so that “tce-setdrive” told me everything was already done, and “backup” told me setup was already done and asked me if I wanted to backup now, to which I answered yes.
When following the exact “import” commands given in the README file (for flwm_topside), the boot process proceeded apparently as expected, but was actually stuck to an endlessly swirling slash, while supposedly loading extensions. I was however able to type Alt+SysRq+E, which left me on TC prompt: there, “mount” told me that no extension was mounted, and /tmp/tcloop was indeed empty.
I also tried with my useless “kmap=…” removed, as well as the “loglevel=3” taken from Isolinux config in the ISO file.
This was in VirtualBox, and then on a real laptop with the same results.
I then started anew, this time import-ing -b xorg-all, and then import -b -f a single file with Xprogs, flwm_topside and wbar in it. This time, the loading of extensions ended correctly (I suppose) but I received this endlessly repeated message (and no prompt):
can't run '/sbin/getty': No such file or directory
For your information, my bootloader is Grub2, and dCore.gz and vmlinuz were taken from the ISO file for RC3.
-
I also tried the instructions in the readme's, using the ISO I burned to CD before.
The tce directory and the sub-directories were created OK on USB stick, and the backup worked (albeit with nothing installed). However when I tried the imports I got the same "could not create symlinks" error messages as before.
edit: I don't get the error messages if I boot from CD with no USB stick tce. I assume in that case the tce directory and packages are being created in memory? Maybe there's some sort of permissions error on the USB stick directories?
-
Re: Pendrives
The only way I could reproduce this error was to attempt to use a ZIP style pendrive. That is a non-linux partition.
Typically the second partition is vfat and it fails. Changing it to ext2 it worked as expected.
However zip type pendrive installs are limited to the smaller pendrives and likely you will run out of space.
-
Re Using the iso
I verified that dCore.gz and vmlinuz inside the iso are the same as the stand-alone versions.
Be sure that you are using a linux partition and that you have not run out of space or co-mingling with Core native backup or other native files.
-
Hi roberts! Thank you for helping me.
My USB pendrive is formatted with a single FAT partition, and a normal DOS partition table, using Gparted. Here's its current status:
[root@archlinux ~]# LANG=C fdisk -l /dev/sdb
Disk /dev/sdb: 16.0 GB, 16008609792 bytes, 31266816 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk label type: dos
Disk identifier: 0x00085ba8
Device Boot Start End Blocks Id System
/dev/sdb1 2048 31266815 15632384 b W95 FAT32
-
As said above, "be sure that you have a linux partition"...
-
My USB pendrive is neither old nor small. It is a 16GB pendrive bought this year. Besides, I only want a single FAT32 partition on it. I don’t need nor want any other partition on it. And it is neither USB-ZIP nor USB-Floppy; it is USB-HDD.
Anyway, I don’t understand: dCore is supposed to run from the "dCore.gz" initramfs. The install media is only meant as storage for the extensions, and FAT32 should be suitable for this purpose. And as can be checked while zero extensions are created (first boot after install, when I still get the prompt), the USB pendrive’s FAT32 partition gets mounted perfectly well, and I can read and write to it normally.
That being said, I miss many extensions in TC5, and I’m willing to try hard to make dCore work, and see if I can replace TC5 with dCore. I can format the pendrive again (even create a new partition table) and start anew.
Keeping in mind I only want a single FAT32 partition, what tools/commands do you propose I run to achieve the right USB media layout for dCore?
-
FWIW, since my last post I've tried using the standard TC 5.0 tc-install to format my USB stick to ext4, then copied over the dCore.gz and vmlinuz files from the dCore ISO and updated the extlinux.conf file to boot dCore. I then deleted the original core.gz and the directories under tce that aren't relevant to dCore. This setup actually booted and I was able to import the packages for the flwm_desktop without apparent errors.
Don't know what the readme instruction to "specify the desktop as a boot parameter" means, if it's something different than the following instruction to add "desktop=flwm_topside at the end of the boot parameter line".
Still got errors (probably because I omitted something above) when I tried to reboot into the desktop. Ended up at the box sign-in.
-
I miss many extensions in TC5..
If you can let us know which ones work in tc-5, we can move them across
-
Juanito: You’re right, and I’ll try and find the time to do that, although the list is quite long :)
But to tell the truth, I’m actually very interested in dCore since I’ve heard about it, and I’d like to see it work.
You may be interested to know that xorg-all is the culprit. I tried with xorg-vesa instead with the same failure:
“can't run '/sbin/getty': No such file or directory”
on and on and on… and no prompt :(
If however I only load Xprogs, flwm_topside, and wbar, boot happens normally, I get a shell prompt, and I can see that the three extensions appear in /tmp/tcloop.
Something must be wrong with xorg or one of its dependencies… Could someone make available to me an xorg-all.sce that is known to work?
-
Confirming what theYinYeti wrote, when I deleted the xorg-all.tce package and removed the name from sceboot.lst on the USB stick it booted to tcbox with no error messages.
-
I'll check into xorg tonight, though it is working fine here and I re-imported it yesterday.
-
Thanks Jason. It's entirely possible that something is wrong with my setup (always my first assumption), but it's interesting that theYinYeti is seeing a similar issue.
edit: What the hey. Re-imported xorg-all without changing anything else and re-booted. Got a desktop! Still had error messages during boot, tcWbarConf button on the Control Panel was dimmed, and wbar had 2 xterminal [sp?] icons, along with an "Import" icon and the usual TC ones. Will try playing with this and see what develops.
-
Sounds normal for this stage of rc.
Two X terms one is UXTerm - unicode the other is normal XTerm
tc-wbarconf is there will investigate why it is greyed on control panel. You can run from Xterm.
Many boot messages are displayed and can be ignored.
-
Umm… I'll try and re-import xorg again, then. I'll do this on the real laptop this time, just in case it might help…
Have a nice day :)
-
No matter how many times I try, xorg-all is still a failure. But if I load-sce xorg-all and startx without rebooting, then xorg does start and work. That is, after I have run “chown root:root /tmp/.X*”.
As soon as I restart, I get that dreaded “can't run '/sbin/getty': No such file or directory” message :(
Even though I don’t see a clear relation between the message and the following “explanation”, do you think it could be because dCore is trying to auto-launch X11 upon auto-login, then X11 fails to start because of the permissions in /tmp/, then login tries again to auto-launch X11, and so on in a loop? What do you think?
And you, how do you handle this problem with permissions in /tmp/ when launching “startx”?
-
What is your permission on /tmp
Should be: drwxrwxrwt root.staff /tmp
No change. Same as Core native.
-
Sorry roberts, I used the wrong word. It's not a permissions issue, but rather a owner issue. /tmp is OK. It is /tmp/.X… (I don't remember the exact name) that is wrongly owned. It is owned by tc:staff whereas startx expect this directory to be owned by root:root.
Just to leave no doubt, my dCore is completely unmodified. I took vmlinuz and dCore.gz from the ISO file, and put them on my USB pendrive. Then I just created an entry for them in my grub.cfg (Grub2).
-
dCore-install is now available. Use the typical import command to fetch.
dCore-install is a command line script to aid in installing dCore to pendives and frugal hard drive.
It it dervived from tc-install so operation is quite similiar. Tested zip, hdd, and frugal.
As with any installation program proceed with caution as specifying the wrong drive/partition can ruin your day.
-
This is a good idea, roberts! I will try this. Thank you.