You are not logged in.
Make sure that your swap file/partition is working?
I had ASCII on sda1 (installed from scratch to straight gpt partition - no special secure boot or anything) for the past several years. To make sure that I still remembered how to install Devuan, I installed Beowulf (a brand new simple - not advanced install), and got it all configured, on a separate partition.
Once I was confident that that freshly-installed Beowulf was working great, on the separate partition, I waited a few days, until I got enough confidence to "risk" trashing my "perfect" existing ASCII install by doing an upgrade-in-place of it to Beowulf, still on sda1. I was very, very pleased with how quickly that upgrade process went, and how correct its results were.
I think that, earlier in this thread, I said that I had installed Beowulf on sda1 from scratch, but tonight, I remembered that, even though I had done that on a separate partition, the Beowulf that I've been using on sda1 USED TO BE ASCII, and was upgraded to Beowulf, preserving its autologin capability.
I don't know if any of that makes any difference to you, but, in case it does, I thought I'd better make it clear, now that I remember how it all happened.
Wow, this is incredible! I looked and looked, but I couldn't find the site that had told me how to do my autologin back when I did it. There appear to be many different choices, depending whether you're using Light-DM or slim, or whatever. CONFUSED WAS I.
THEN, I took a chance and looked in my own saved installation notes from back then. And I found the instructions that I had saved, to make it easier to do autologins in the future:
LightDM configuration file is found at /etc/lightdm/lightdm.conf. Making a backup of the original configuration file is recommended.
Enable autologin
Look up this line in lightdm configuration file, UNCOMMENT IT and add your username.
CHANGE:
#autologin-user=
TO:
autologin-user=computerbob
NOTE: #autologin-user APPEARS IN TWO DIFFERENT PLACES IN THE LIGHTDM.CONF FILE.
THE ONLY ONE THAT NEEDS TO BE CHANGED IS THE ONE IN WHICH THE EQUALS SIGN APPEARS IMMEDIATELY AFTER #autologin-user=
THE OTHER PLACE THAT IT APPEARS, THE EQUALS SIGN APPEARS AFTER A SPACE, LIKE THIS #autologin-user =
DON'T EDIT THAT LINE - ONLY EDIT THE LINE IN WHICH THE EQUALS SIGN APPEARS IMMEDIATELY AFTER #autologin-user
I'm so glad that my HD has a lot better (and longer) memory than I do!
I still want to know what you did to set up autologin so that I fully understand the cause of the problem, in case someone else does the same.
I'm sorry I didn't remember to figure that out yet. I'm using Xfce, with whatever the default login manager is, but I first set it up years ago, in ASCII, and it was automatically upgraded to work when I finally got up the motivation and courage to upgrade ASCII to Beowulf a few weeks ago.
I promise that I will figure out what I did, and where I did it, to make autologin work.
Glad you got it working Computer Bob.
Thank you guys!!!! Seriously, YOU GUYS got it working. I just followed instructions.
IIRC, the whole "squashing" process took several minutes, with virtually no feedback, because the gui feedback progress, back-and-forth bar slowed way, way, down, then stopped altogether, even though (I figured out later) it was actually still "silently" working for several minutes -- while the command-line progress bar only updated two or three times during the whole process. But I'm really glad that they worked, even though, at the time, I suspected that they had locked up.
Thank you, again! And thank you for the link -- I look forward to learning a lot more about RefractaSnapshot, once enough of my brain cells have recovered!
YES! I added noautologin to the end of the boot line, and then rebooted!
SUCCESS - except that my data partition didn't get mounted
so that meant that I had no email or browser, UNTIL I eventually figured out how to mount the data partition (/dev/sda6) as /media/data, because that's where my email and browser profiles are stored.
But I got it figured out, as proven by the fact that I'm writing this very first forum response
FROM MY BEOWULF SYSTEM, FULLY CONFIGURED, RUNNING ON MY USB STICK!
I don't know if it's a real word or not, but you have all been patient, knowlegable, understanding, genii (the plural of "genius," in my mind) throughout this entire, lengthy, time-consuming thread.
It has been a REAL HONOR to work with you. Thank you very, very much. You have no idea how much good it did me to help in my own small way. And I hope that none of you ever have a stroke, and learn, first-hand, how much you've all done for me.
But, please know, in your hearts, from my words, that you've done a tremendous thing for me.
I think there's supposed to be a way to use my WORKING USB stick to INSTALL my system onto a hard drive, but I'm not even going to worry about that right now.
THANK YOU, THANK YOU, THANK YOU!
If you have any unanswered questions, or things for me to test, I'll be happy to keep replying!
It is SO CLOSE!!!!
It was a scary process. I remember that it seemed to take a lot longer than in the past, to run the snapshot-creation program.
The "squashing" section itself probably took ten minutes, while the little pop-up gui feedback window never showed anything except that it reported 0% complete the whole time.
I think it may have been at the same time that the command line showed a progress bar that stayed at 0% for a long time, then jumped to 20,xxx for a long time, then 40,xxx for a long time, then completed without any further updates.
Then, I think the gui window disappeared for awhile, before returning to tell me that it was all done!
There was no longer any /etc/home/ folder. I found the iso and its SHA file in /media/data/snapshot/
Copying the iso to sdb took much longer than it had before. I considered that to be a GOOD sign, except that the only way I could tell that it was doing anything was by the quickly-flashing (not slowly-flashing) light on the USB stick.
Then, my USB stick BOOTED UP to my complete, auto-logged-in desktop (the autolog-in is intentional). But for only a few seconds.
Then it dropped back to a CLI screen with messages at the bottom, for only about one second, before returning to my desktop, with a login screen, autofilled with my username, for about 3 seconds.
Then back to the CLI screen for about one second, back to the desktop for about 3 seconds. After the 5th of 6th infinite circle or activities, I physically shut down my computer, and booted back to the HD again. That's how I'm typing this right now.
MAN, THIS IS SO COOL! IT IS SO CLOSE TO WORKING!!!!!!!!!
Messages received, guys! And some of them actually UNDERSTOOD.
I have printed them out, and will start to work on them, as a to-do list.
I'll replace the older kernel with a newer fallback one, as you suggested.
"Perhaps you haven't created the snapshot correctly." IIRC, alll I did was click on the the default option to create the snapshot, and then it did everything else itself. Did I miss something? Do I need to run something else in addition to what I ran, to create the snapshots?
OK, I just removed/installed the kernels, exactly as you suggested. In Synaptic, that caused the following two packages to appear as "Installed (auto removable)":
irqbalance
runit helper
I have not removed those two packages (yet?), just in case you wanted me to NOT do that.
I will reboot and continue on to creating a new snapshot (hopefully, tomorrow). An entire day today, of trying to help in a domestic violence situation, and dealing with someone who just had a stroke, really tired me out... but I'll still be here again, as soon as I can.
Now, it's a few minutes later. I just rebooted. A quick Catfish File Search of the whole File System found 14 items with "squashfs" in their names. Five of them were folders, and 9 of them were files. (???)
Good night! And Thanks.
I lost track of what we're doing a few days ago. Right now, I'm just a monkey with a keyboard.
I'm sure glad that you guys know what you're doing.
computerbob@robinson:~$ uname -a
Linux robinson 4.19.0-14-amd64 #1 SMP Debian 4.19.171-2 (2021-01-30) x86_64 GNU/Linux
computerbob@robinson:~$
computerbob@robinson:/media/computerbob/liveiso$ ls -lh /media/computerbob/liveiso/live
total 42M
-r--r--r-- 1 computerbob computerbob 37M Mar 15 16:04 initrd.img
-r--r--r-- 1 computerbob computerbob 179K Dec 12 2019 memtest
-r--r--r-- 1 computerbob computerbob 5.1M Mar 15 16:01 vmlinuz
computerbob@robinson:/media/computerbob/liveiso$
ralph.ronnquist wrote:This kind of presents the same hardware picture as your post #34 execpt that it has sda4 instead of sda6.
That difference is probably the cause of your problem... and now we need @fsmithred to draw conclusions
I don't see why that would matter. It should be booting the live system, not a disk partition. But it does lead me to a question. What's in the boot command?
Bob, at the boot menu for the live-usb, press TAB and see what it shows at the bottom of the screen. It should be
/live/vmlinuz initrd=/live/initrd.img boot=live username=computerbob
I can boot to initramfs prompt if I remove 'boot=live' from the boot command. I'm wondering if your boot menu got screwed up and lost that piece.
I'll try to do that now, and then report back in this same message. Please hold.... (Muzak: la, la, la, dum-de-dum-dum... Please continue to hold. Your call is very important to us. Estimated wait time is approximately 38 hours...)
I will try to copy what I see in the blurry photo that I just took after pressing TAB at the boot menu:
/live/vmlinuz initrd=/live/initrd.img boot=live username=computerbob _
I suspect that that may be what you said it might be, but I'm too tired tonight (it's been a long, long but rewarding day) to compare the two lines myselt.
That's good information. Only the s* were interesting and in particular the sda* and sdb* entries which are adapters for SATA devices and their partitions. Apparently there is one "drive", sda, with 4 partitions and another, sdb, with 1 partition.
This kind of presents the same hardware picture as your post #34 execpt that it has sda4 instead of sda6.
That difference is probably the cause of your problem... and now we need @fsmithred to draw conclusions
You are correct about sda and sdb. The sda drive with 4 partitions is my HD, my sdb USB stick was totally unallocated space until I copied the Snapshot to it -- I didn't intentionally create any partitions on it - maybe that was created by the copying process?
GParted currently shows me the following on my sda:
/dev/sda1 /
/dev/sda2 /home
/dev/sda3 /linux-swap
unallocated 9.69 GB
/dev/sda6 /media/data (102.54 GB of user data)
They're not especially sharp photos -- the post-its on my monitor are in sharper forcus that the screen itself -- but, as best as I can make out, here are the things that start with "s." It might take me awhile, because I think they are displayed horizontally across 3 of the 4 photos...
sda
sda1
sda2
sda3
sda4
sdb
sdb1
stdin?
sg0?
sg1?
sg2?
tty0
tty1
tty10
tty11
tty12
tty13
tty14
tty15
tty16
tty17
tty18
tty19
tty2
tty20 through tty29
tty3
tty 30 through 39
tty4
tty 40 through 49
urandom
ucs?
ucs1?
ucsa?
ucsa1?
vcsa?
vcsa1?
vcsu?
vcsu1?
vga_arbiter
zero
ztdin?
ztdout?
This was one of the most time-consuming, hardest things I've tried to do in a long, long time -- remembering one number at a time, going back and forth from the forum to blurry photos in GIMP. Whew! I sure hope I could see enough info to help you!
@ralph,
I just took your advice, but I don't know any way to attach my 4 photos to this forum message or to the private forum message that I sent you, and I don't have any photo-displaying service to put them on. ;(
Are you wanting me to go back to the "crashed" (initramfs) prompt and copy down the names of the files, starting with the ones that start with "s"?
I'm about to leave to go do some ministry, but I would be able to do it when I return.
I don't know of any way to copy/paste it all here, but running
ls /dev/
at the (initramfs) prompt showed me 129 files, with several different names.
Ok so you need to remove /etc/initramfs-tools/conf.d/driver-policy "driver-policy" file.
Then
# update-initramfs -u
Then make a new snapshot.
I had hoped to be writing this from my system, after booting from my USB stick.
I was pretty hopeful, because the whole iso creation and copying it to my USB stick seemed noticeably faster than it had been previously.
But I still I got the exact same frozen-after-the "random:crng init done" line, followed by a long pause, followed by the whole "BOOT FAILED" message -- which still complains that it can't find a bootable system...
I'm really sorry guys. With it working for other people now, it's starting to clearly look and feel like I must be the problem.
computerbob@robinson:~$ ls -lh /boot
total 57M
-rw-r--r-- 1 root root 202K Jan 30 04:35 config-4.19.0-14-amd64
-rw-r--r-- 1 root root 183K Dec 17 07:51 config-4.9.0-14-amd64
drwxr-xr-x 5 root root 4.0K Mar 10 10:30 grub
-rw-r--r-- 1 root root 37M Mar 15 15:58 initrd.img-4.19.0-14-amd64
-rw-r--r-- 1 root root 4.7M Dec 18 13:22 initrd.img-4.9.0-14-amd64
-rw-r--r-- 1 root root 3.3M Jan 30 04:35 System.map-4.19.0-14-amd64
-rw-r--r-- 1 root root 3.1M Dec 17 07:51 System.map-4.9.0-14-amd64
-rw-r--r-- 1 root root 5.1M Jan 30 04:35 vmlinuz-4.19.0-14-amd64
-rw-r--r-- 1 root root 4.1M Dec 17 07:51 vmlinuz-4.9.0-14-amd64
I suppose initrd.img will vary from system to system but my initrd.img is 30M, the system that had limited drivers was around 19M.
Is the file /etc/initramfs-tools/conf.d/driver-policy present ?
Yes, its contents are as follows:
# Driver inclusion policy selected during installation
# Note: this setting overrides the value set in the file
# /etc/initramfs-tools/initramfs.conf
MODULES=dep
What does Computer Bobs /etc/initramfs-tools/initramfs.conf look like ?
#
# initramfs.conf
# Configuration file for mkinitramfs(8). See initramfs.conf(5).
#
# Note that configuration options from this file can be overridden
# by config files in the /etc/initramfs-tools/conf.d directory.
#
# MODULES: [ most | netboot | dep | list ]
#
# most - Add most filesystem and all harddrive drivers.
#
# dep - Try and guess which modules to load.
#
# netboot - Add the base modules, network modules, but skip block devices.
#
# list - Only include modules from the 'additional modules' list
#
MODULES=most
#
# BUSYBOX: [ y | n | auto ]
#
# Use busybox shell and utilities. If set to n, klibc utilities will be used.
# If set to auto (or unset), busybox will be used if installed and klibc will
# be used otherwise.
#
BUSYBOX=auto
#
# KEYMAP: [ y | n ]
#
# Load a keymap during the initramfs stage.
#
KEYMAP=n
#
# COMPRESS: [ gzip | bzip2 | lz4 | lzma | lzop | xz ]
#
COMPRESS=gzip
#
# NFS Section of the config.
#
#
# DEVICE: ...
#
# Specify a specific network interface, like eth0
# Overridden by optional ip= or BOOTIF= bootarg
#
DEVICE=
#
# NFSROOT: [ auto | HOST:MOUNT ]
#
NFSROOT=auto
#
# RUNSIZE: ...
#
# The size of the /run tmpfs mount point, like 256M or 10%
# Overridden by optional initramfs.runsize= bootarg
#
RUNSIZE=10%
Just in case this information is helpful, too:
computerbob@robinson:~$ uname -a
Linux robinson 4.19.0-14-amd64 #1 SMP Debian 4.19.171-2 (2021-01-30) x86_64 GNU/Linux
computerbob@robinson:~$
@ComputerBob:
When you return to this, here's a simple check to see if your problem is the same as dice's. Adjust the file name if you have a different kernel.ls -lh /boot/initrd.img-4.19.0-14-amd64
Do you recall if you did an expert install and chose to only have drivers targeted for your hardware in the initramfs? If you don't recall, that's fine. The output of the command above will tell us. It'll either be big or small. (I want the number, please.)
Even before I run that command, in Thunar, I can see one /grub/ folder and eight files inside the /boot/ folder, but I'm guessing that the ls command itself might show you those anyway:
computerbob@robinson:~$ ls -lh /boot/initrd.img-4.19.0-14-amd64
-rw-r--r-- 1 root root 24M Mar 13 13:12 /boot/initrd.img-4.19.0-14-amd64
computerbob@robinson:~$
Since I really don't know what I'm doing, here's a different version of the same ls command, to show the entire contents of that folder:
computerbob@robinson:~$ ls -lh /boot
total 44M
-rw-r--r-- 1 root root 202K Jan 30 04:35 config-4.19.0-14-amd64
-rw-r--r-- 1 root root 183K Dec 17 07:51 config-4.9.0-14-amd64
drwxr-xr-x 5 root root 4.0K Mar 10 10:30 grub
-rw-r--r-- 1 root root 24M Mar 13 13:12 initrd.img-4.19.0-14-amd64
-rw-r--r-- 1 root root 4.7M Dec 18 13:22 initrd.img-4.9.0-14-amd64
-rw-r--r-- 1 root root 3.3M Jan 30 04:35 System.map-4.19.0-14-amd64
-rw-r--r-- 1 root root 3.1M Dec 17 07:51 System.map-4.9.0-14-amd64
-rw-r--r-- 1 root root 5.1M Jan 30 04:35 vmlinuz-4.19.0-14-amd64
-rw-r--r-- 1 root root 4.1M Dec 17 07:51 vmlinuz-4.9.0-14-amd64
computerbob@robinson:~$
For the first time in many years, I used the first (simpler) install option, instead of the more-detailed expert install option, installing Beowulf onto partitions that I had already created in GParted. I don't remember if that gave me the option to install only specific drivers or all drivers. However, if it DID give me that option, I know that -- in the past - - I've always chosen to install ALL drivers.
My wife and I have to leave to go to a lengthy appointment in 20 minutes. Before I leave, or after I get back, I may or may not have time to do any more troubleshooting (or dumb typing, depending on how you see it) today.
I don't understand the symlinks at all. When you looked at the mounted usb with ls, it showed the right things.
You can mount the iso as root and look at it that way.
mount /home/snapshot/snapshot.iso /mnt ls /mnt ls /mnt/live ls /mnt/isolinux # and when you're done looking around in there umount /mnt
You should see /live /isolinux /pkglist_whatever. Look inside /live and you should see initrd.img vmlinuz and filesystem.squashfs. Inside /isolinux should be live.cfg (the boot menu) and a bunch of other files that are the same ones in /usr/lib/refractasnapshot/iso/isolinux/.
Edit: you could do the same with the mounted usb and look inside the directories.
ls /media/computerbob/liveiso/live # and so on.
Yes, everything I saw was there and, and far as I could tell, it all looked like real files. I haven't checked the USB stick yet. I may or may be able to figure out how to do that. Is there ANY possibility that there are snapshot FILES on my HD, but they're somehow converted to symlinks on my USB stick when I'm copying them there? So my USB stick boots (linked to existing files on my HD), until it wants something that it can't find. I'm just shooting in the dark here, so just ignore me if my ideas are not helpful.
NEVER MIND - I figured out how to mount my USB stick and view its various contents, using terminal commands. Everything I've seen there looks like REAL files, not symlinks. Man, this stuff is a real challenge to me!