Hello everyone, I am trying to compile a old kernel and it compiles fine. I used: time make -j$(nproc) and this is a-ok and works. make modules_install works fine. Now when I try to install the newly compiled kernel I get this error below. Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64) it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for more information. and this file shows me what is below: ============================================================== DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64) Tue Jun 22 22:30:14 EDT 2021 /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. =============================================================== I can use all the help I can get right now. Anyone ever experience this ? If so is is there a fix or workaround ? What causes this error ? I have multiple kernels on this ancient box and would really prefer not to have to wipe everything and reinstall from scratch. The only thing I did recently was install Virtualbox and Wireshark. Both work no issues. If anyone has any pointers would be most welcome and much appreciated. Thanks - Aruna
I am trying to compile a old kernel and it compiles fine. I used: time make -j$(nproc) and this is a-ok and works. make modules_install works fine. Now when I try to install the newly compiled kernel I get this error below.
Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)
Just to be clear: This is what you get when running "sudo make install" ?
it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for more information. and this file shows me what is below:
============================================================== DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64) Tue Jun 22 22:30:14 EDT 2021 /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. ===============================================================
What are the contents of this Makefile (line 167)? Looks like it's expecting a "Makefile.normal" that it can't find - I don't know what this file is or where it would come from. All the best, -- Thomas
On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen <t@laumann.xyz> wrote:
I am trying to compile a old kernel and it compiles fine. I used: time make -j$(nproc) and this is a-ok and works. make modules_install works fine. Now when I try to install the newly compiled kernel I get this error below.
Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)
Just to be clear: This is what you get when running "sudo make install" ?
Yes, I was getting that error when running "sudo make install". I have manually removed all instances of VirtualBox and any guest additions and that warning went away. But now I have another problem. After compilation grub menu shows the new kernel. When I select the new kernel it goes through all the usual steps of booting then hangs at the graphical login screen where I no longer have a mouse and Ctrl+Alt+F1 does not work either . I have given below the output of: sudo time make install 2>&1 | tee make_install.txt ============================================================== sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \ System.map "/boot" run-parts: executing /etc/kernel/postinst.d/apt-auto-removal 3.18.2-testing /boot/vmlinuz-3.18.2-testing run-parts: executing /etc/kernel/postinst.d/initramfs-tools 3.18.2-testing /boot/vmlinuz-3.18.2-testing update-initramfs: Generating /boot/initrd.img-3.18.2-testing W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for module r8169 run-parts: executing /etc/kernel/postinst.d/zz-update-grub 3.18.2-testing /boot/vmlinuz-3.18.2-testing Generating grub configuration file ... Found background image: /usr/share/images/desktop-base/desktop-grub.png Found linux image: /boot/vmlinuz-3.18.2-testing Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.18.2-testing.old Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.16.0-4-amd64 Found initrd image: /boot/initrd.img-3.16.0-4-amd64 Found memtest86+ image: /boot/memtest86+.bin Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin Found Windows 7 (loader) on /dev/sda1 Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1 Found Debian GNU/Linux (6.0.5) on /dev/sdb3 Found Debian GNU/Linux (8.0) on /dev/sdb5 Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6 done 1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k 251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps ============================================================== Like I said I have a lot of old kernels lying around this box :-)
it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for more information. and this file shows me what is below:
============================================================== DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64) Tue Jun 22 22:30:14 EDT 2021 /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. /bin/sh: 1: test: -I./arch/x86/include: unexpected operator Using standalone build system. Makefile:167: Makefile.normal: No such file or directory make[2]: *** No rule to make target 'Makefile.normal'. Stop. ===============================================================
What are the contents of this Makefile (line 167)? Looks like it's expecting a "Makefile.normal" that it can't find - I don't know what this file is or where it would come from.
The contents of the Makefile in ./arch/x86/include at line 167: 166 - # prevent gcc from generating any FP code by mistake 167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow 168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,) 170 - KBUILD_CFLAGS += $(mflags-y) 171 - KBUILD_AFLAGS += $(mflags-y)
All the best, -- Thomas
You have pointed me in the right direction so thank you. Ten plus years I have been dabbling with Linux and I just learnt something new. I had no idea about this: aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh /etc/kernel/postinst.d total 20K drwxr-xr-x 2 root root 4.0K Jun 23 06:18 . drwxr-xr-x 7 root root 4.0K Mar 31 00:45 .. -rwxr-xr-x 1 root root 2.8K Apr 13 2015 apt-auto-removal -rwxr-xr-x 1 root root 858 Mar 1 2015 initramfs-tools -rwxr-xr-x 1 root root 641 Dec 14 2015 zz-update-grub Thank you Thomas I will keep trying to fix the no mouse issue now. Aruna
On 6/24/21, Aruna Hewapathirane <aruna.hewapathirane@gmail.com> wrote:
On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen <t@laumann.xyz> wrote:
I am trying to compile a old kernel and it compiles fine. I used: time make -j$(nproc) and this is a-ok and works. make modules_install works fine. Now when I try to install the newly compiled kernel I get this error below.
Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)
Just to be clear: This is what you get when running "sudo make install" ?
Yes, I was getting that error when running "sudo make install". I have manually removed all instances of VirtualBox and any guest additions and that warning went away. But now I have another problem. After compilation grub menu shows the new kernel. When I select the new kernel it goes through all the usual steps of booting then hangs at the graphical login screen where I no longer have a mouse and Ctrl+Alt+F1 does not work either .
Hi from a perennial lurker here! Disclaimer up front: My apologies if you're already doing this step. In my case, my experience is with Debian doing similar versions of this over the last year or so. I tripped over that I was mismatching e.g. my initrd.img and vmlinuz instances in an attempt at separating out all things "/boot" while running LILO instead of GRUB2 for 14 or so different bootable partitions. My stumbled upon fix was to run "update-initramfs -u" then double check to make sure that that latest update matched the location LILO was using to call up those partition-specific files. As an example, I have a "lilo" subdirectory under /boot then each operating system has its own deeper subdirectory. Those subdirectories hold copies of each system's initrd.img and vmlinuz. Ever since I started very carefully double checking everything after each update that triggered LILO to run, I haven't been locked out since, not once. Prior to this (really just a few weeks ago), I faced every possible form of lockout that you're describing plus occasionally also hitting that blank screen where the login prompt doesn't even bother to appear. The experience has been so empowering that I'm preparing to test drive syslinux, too, something I've actually "feared" until a couple weeks ago. :) Hope this helps someone at some point... somewhere.. somehow. Cindy :) -- * runs with birdseed *
On Thu, Jun 24, 2021 at 12:13 PM Cindy Sue Causey <butterflybytes@gmail.com> wrote:
On 6/24/21, Aruna Hewapathirane <aruna.hewapathirane@gmail.com> wrote:
On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen <t@laumann.xyz> wrote:
I am trying to compile a old kernel and it compiles fine. I used: time make -j$(nproc) and this is a-ok and works. make modules_install works fine. Now when I try to install the newly compiled kernel I get this error below.
Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)
Just to be clear: This is what you get when running "sudo make install" ?
Yes, I was getting that error when running "sudo make install". I have manually removed all instances of VirtualBox and any guest additions and that warning went away. But now I have another problem. After compilation grub menu shows the new kernel. When I select the new kernel it goes through all the usual steps of booting then hangs at the graphical login screen where I no longer have a mouse and Ctrl+Alt+F1 does not work either .
Hi from a perennial lurker here! Disclaimer up front: My apologies if you're already doing this step.
out all things "/boot" while running LILO instead of GRUB2 for 14 or so different bootable partitions.
Hi, I thought 'I' was the only person insane enough to have multiple partitions that run Linux ? 14 or so partitions ? Yikes ! Let's see how many I have managed to create over the years : aruna@debian:~$ lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT sda 8:0 0 232.9G 0 disk ├─sda1 8:1 0 100M 0 part ├─sda2 8:2 0 212.2G 0 part /media/aruna/OS ├─sda3 8:3 0 1K 0 part ├─sda5 8:5 0 19.7G 0 part / └─sda6 8:6 0 903M 0 part [SWAP] sdb 8:16 0 931.5G 0 disk ├─sdb1 8:17 0 250G 0 part /media/aruna/Ubuntu 12.04.4 L ├─sdb2 8:18 0 250G 0 part /media/aruna/kernel-stuff ├─sdb3 8:19 0 16.8G 0 part /media/aruna/back-up ├─sdb4 8:20 0 1K 0 part ├─sdb5 8:21 0 204G 0 part /media/aruna/3c8fce17-5948-43e7-9662-921e65c7faa7 ├─sdb6 8:22 0 196.9G 0 part /media/aruna/200f9328-ae6b-4fe9-9d94-24ffd391bbe9 ├─sdb7 8:23 0 6.2G 0 part └─sdb8 8:24 0 7.7G 0 part sdh 8:112 1 3.7G 0 disk ├─sdh1 8:113 1 2.4G 0 part /media/aruna/d-live 10.10.0 ma amd64 └─sdh2 8:114 1 2.5M 0 part sr0 11:0 1 1024M 0 rom Darn! You beat me by one partition but it is a close second eh ? Can't count the sdh that has the Debian 10 live CD I just plugged in :-)
In my case, my experience is with Debian doing similar versions of this over the last year or so. I tripped over that I was mismatching e.g. my initrd.img and vmlinuz instances in an attempt at separating
Believe me I am feeling the pain you must have felt at that time right now. And this is really interesting what you have shared, makes me wonder now.
My stumbled upon fix was to run "update-initramfs -u" then double check to make sure that that latest update matched the location LILO was using to call up those partition-specific files. As an example, I have a "lilo" subdirectory under /boot then each operating system has its own deeper subdirectory. Those subdirectories hold copies of each system's initrd.img and vmlinuz.
I have always used GRUB2 so now I have to read about "lilo" I guess. Thanks.
Ever since I started very carefully double checking everything after each update that triggered LILO to run, I haven't been locked out since, not once. Prior to this (really just a few weeks ago), I faced every possible form of lockout that you're describing plus occasionally also hitting that blank screen where the login prompt doesn't even bother to appear. The experience has been so empowering that I'm preparing to test drive syslinux, too, something I've actually "feared" until a couple weeks ago. :)
You 'feared' syslinux ? There was a time long years ago I had accidentally opened up a shell/terminal and I just sat there for a good 30 minutes unsure what the hell do I do now ? I was too afraid to x-out and close that window because I had no idea what it would or would not do. And I had no clue what to type in that bash shell either. Those were the good old days :-)
Hope this helps someone at some point... somewhere.. somehow.
It has most definitely helped me. I will be back with more questions as I proceed to try and troubleshoot this mess I have created. Thank you very much for sharing - Aruna
Cindy :) -- * runs with birdseed *
_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Yes, I was getting that error when running "sudo make install". I have manually removed all instances of VirtualBox and any guest additions and that warning went away.
Interesting...
When I select the new kernel it goes through all the usual steps of booting then hangs at the graphical login screen where I no longer have a mouse and Ctrl+Alt+F1 does not work either .
Is there a way you could increase the kernel log level to debug, and capture the logs during boot? That might be informative.
I have given below the output of: sudo time make install 2>&1 | tee make_install.txt
============================================================== sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \ System.map "/boot" run-parts: executing /etc/kernel/postinst.d/apt-auto-removal 3.18.2-testing /boot/vmlinuz-3.18.2-testing run-parts: executing /etc/kernel/postinst.d/initramfs-tools 3.18.2-testing /boot/vmlinuz-3.18.2-testing update-initramfs: Generating /boot/initrd.img-3.18.2-testing W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for module r8169 run-parts: executing /etc/kernel/postinst.d/zz-update-grub 3.18.2-testing /boot/vmlinuz-3.18.2-testing Generating grub configuration file ... Found background image: /usr/share/images/desktop-base/desktop-grub.png Found linux image: /boot/vmlinuz-3.18.2-testing Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.18.2-testing.old Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.16.0-4-amd64 Found initrd image: /boot/initrd.img-3.16.0-4-amd64 Found memtest86+ image: /boot/memtest86+.bin Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin Found Windows 7 (loader) on /dev/sda1 Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1 Found Debian GNU/Linux (6.0.5) on /dev/sdb3 Found Debian GNU/Linux (8.0) on /dev/sdb5 Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6 done 1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k 251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps
==============================================================
Like I said I have a lot of old kernels lying around this box :-)
And Windowses :-) But nothing looks particularly erroneous about the output.
The contents of the Makefile in ./arch/x86/include at line 167:
166 - # prevent gcc from generating any FP code by mistake 167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow 168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,)
Interesting, I see no mention of "Makefile.normal" here. But maybe that's because you removed whatever required it?
You have pointed me in the right direction so thank you. Ten plus years I have been dabbling with Linux and I just learnt something new. I had no idea about this:
aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh /etc/kernel/postinst.d total 20K drwxr-xr-x 2 root root 4.0K Jun 23 06:18 . drwxr-xr-x 7 root root 4.0K Mar 31 00:45 .. -rwxr-xr-x 1 root root 2.8K Apr 13 2015 apt-auto-removal -rwxr-xr-x 1 root root 858 Mar 1 2015 initramfs-tools -rwxr-xr-x 1 root root 641 Dec 14 2015 zz-update-grub
Thank you Thomas I will keep trying to fix the no mouse issue now.
No problem, I wasn't sure if I was helping :-) -- Thomas
On Thu, Jun 24, 2021 at 1:29 PM Thomas Bracht Laumann Jespersen <t@laumann.xyz> wrote:
Yes, I was getting that error when running "sudo make install". I have manually removed all instances of VirtualBox and any guest additions and that warning went away.
Interesting...
Interesting ? Don't forget I compiled and installed via make install, so I had to manually remove the following entries: /boot/vmlinuz*KERNEL-VERSION* /boot/initrd*KERNEL-VERSION* /boot/System-map*KERNEL-VERSION* /boot/config-*KERNEL-VERSION* /lib/modules/*KERNEL-VERSION*/ /var/lib/initramfs/*KERNEL-VERSION*/ Then update the grub configuration: sudo update-grub2 Then look for any loose ends using 'locate': locate -b -e *dkms 3.18.2-* You get the idea. Was definitely not fun :-)
When I select the new kernel it goes through all the usual steps of booting then hangs at the graphical login screen where I no longer have a mouse and Ctrl+Alt+F1 does not work either .
Is there a way you could increase the kernel log level to debug, and capture the logs during boot? That might be informative.
Hmm... should be possible. Never needed to do this before so another learning experience for me here. I will get back to you on this.
I have given below the output of: sudo time make install 2>&1 | tee make_install.txt
============================================================== sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \ System.map "/boot" run-parts: executing /etc/kernel/postinst.d/apt-auto-removal 3.18.2-testing /boot/vmlinuz-3.18.2-testing run-parts: executing /etc/kernel/postinst.d/initramfs-tools 3.18.2-testing /boot/vmlinuz-3.18.2-testing update-initramfs: Generating /boot/initrd.img-3.18.2-testing W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for module r8169 W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for module r8169 run-parts: executing /etc/kernel/postinst.d/zz-update-grub 3.18.2-testing /boot/vmlinuz-3.18.2-testing Generating grub configuration file ... Found background image: /usr/share/images/desktop-base/desktop-grub.png Found linux image: /boot/vmlinuz-3.18.2-testing Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.18.2-testing.old Found initrd image: /boot/initrd.img-3.18.2-testing Found linux image: /boot/vmlinuz-3.16.0-4-amd64 Found initrd image: /boot/initrd.img-3.16.0-4-amd64 Found memtest86+ image: /boot/memtest86+.bin Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin Found Windows 7 (loader) on /dev/sda1 Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1 Found Debian GNU/Linux (6.0.5) on /dev/sdb3 Found Debian GNU/Linux (8.0) on /dev/sdb5 Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6 done 1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k 251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps
==============================================================
Like I said I have a lot of old kernels lying around this box :-)
And Windowses :-) But nothing looks particularly erroneous about the output.
I used to work from home as a remote agent for a call center and they had a policy we had to use Windows no exceptions. I tried to convert them to Linux and you all know how that went :-)
The contents of the Makefile in ./arch/x86/include at line 167:
166 - # prevent gcc from generating any FP code by mistake 167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow 168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,)
Interesting, I see no mention of "Makefile.normal" here. But maybe that's because you removed whatever required it?
Could well be ... but 'locate' shows me this: aruna@debian:~$ locate Makefile.normal /usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal /var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal Does that help or make any sense ?
You have pointed me in the right direction so thank you. Ten plus years I have been dabbling with Linux and I just learnt something new. I had no idea about this:
aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh /etc/kernel/postinst.d total 20K drwxr-xr-x 2 root root 4.0K Jun 23 06:18 . drwxr-xr-x 7 root root 4.0K Mar 31 00:45 .. -rwxr-xr-x 1 root root 2.8K Apr 13 2015 apt-auto-removal -rwxr-xr-x 1 root root 858 Mar 1 2015 initramfs-tools -rwxr-xr-x 1 root root 641 Dec 14 2015 zz-update-grub
Thank you Thomas I will keep trying to fix the no mouse issue now.
No problem, I wasn't sure if I was helping :-)
Yes sir you most certainly did help.
-- Thomas
On Thu, 24 Jun 2021 14:19:04 -0400, Aruna Hewapathirane said:
Could well be ... but 'locate' shows me this:
aruna@debian:~$ locate Makefile.normal /usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal /var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal
Note that 'locate' doesn't actually walk through the file system, it greps through a file created by 'updatedb' which usually gets run from a crontab entry. As a result, 'locate' can report the existence of files that have actually been removed after the last updatedb run.
On Thu, Jun 24, 2021 at 7:09 PM Valdis Klētnieks <valdis.kletnieks@vt.edu> wrote:
On Thu, 24 Jun 2021 14:19:04 -0400, Aruna Hewapathirane said:
Could well be ... but 'locate' shows me this:
aruna@debian:~$ locate Makefile.normal /usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal /var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal
Note that 'locate' doesn't actually walk through the file system, it greps through a file created by 'updatedb' which usually gets run from a crontab entry. As a result, 'locate' can report the existence of files that have actually been removed after the last updatedb run.
Oh dear... no wonder locate was lightning fast and I was thinking even for linux this is crazy fast speed. Now I know... Thank you Valdis and any thoughts on how I can get my mouse to work at the graphical login screen again ? make j$(nproc) is ok make modules_install is ok make install is ok but after selecting the new kernel at the GRUB2 menu it hangs at the graphical login and I am stumped. Soo... that is why 'locate' is so damn fast :-)
On Thu, 24 Jun 2021 19:36:08 -0400, Aruna Hewapathirane said:
but after selecting the new kernel at the GRUB2 menu it hangs at the graphical login and I am stumped.
Does control-alt-F2, -F3, -F4 get you line mode consoles? If so, you should be able to login, and get dmesg, the X server log, and the logs for whatever greeter program runs for graphical login. If you can't get a line mode console, check if it's answering ping, and if it's answering any tcp or udp based services (if you have any enabled). Another thing to look at is netconsole to send kernel messages to a syslog server running on another machine (or netcat at the destination if it doesn't have syslog running),
On Thu, Jun 24, 2021 at 8:51 PM Valdis Klētnieks <valdis.kletnieks@vt.edu> wrote:
On Thu, 24 Jun 2021 19:36:08 -0400, Aruna Hewapathirane said:
but after selecting the new kernel at the GRUB2 menu it hangs at the graphical login and I am stumped.
Does control-alt-F2, -F3, -F4 get you line mode consoles? If so, you should be able to login, and get dmesg, the X server log, and the logs for whatever greeter program runs for graphical login.
No my keyboard is locked not responsive just the numLock led is on, so unable to control-alt-F2 :-(
If you can't get a line mode console, check if it's answering ping, and if it's answering any tcp or udp based services (if you have any enabled). Another thing to look at is netconsole to send kernel messages to a syslog server running on another machine (or netcat at the destination if it doesn't have syslog running),
I found a 500Gb ssd inside one of the old machines and I started from scratch. Pulled out the two cables from the harddisks and plugged in the ssd, installed debian-10.10.0-amd64-netinst.iso then tried to compile kernel 5.12.13 from kernel.org and voila et success. And under 13 minutes to full compile. Also compiled linux-next-next-20210625 no issues, everything is good again. Someday I want to try and figure out what is going on with the original system. Oh, and Thomas Syslog tells me : Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs' Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules. Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state. Does that make sense to you or anyone ? Anyways things are all good. I just have to pull out cables and hook up the ssd when I feel like compiling a kernel, which I can live with :-) Cindy Sue Causey I was wondering is there a 'reason' you are sticking with 'lilo' and not moving to grub2 ? Thank you everyone for all the help and support - Aruna
On Fri, 25 Jun 2021 16:05:45 -0400, Aruna Hewapathirane said:
Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs' Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules. Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
Does that make sense to you or anyone ?
For reasons understandable only to systemd, it tried to modprobe vmhgfs and didn't find it in the initramfs. Does this configuration actually *need* a vmware shared folder in order to boot? If so, you'll need to tell your dracut/ mkinitramfs/whatever command to include it. If not, you'll need to tell it to not look for it. (Particularly annoying - dracut will include that systemd unit even if you have a kernel that doesn't need any modules to boot because all the requisites are built-in, you said --no-kernel to not include kernel modules in the initramfs, and you gave it -o "kernel-modules kernel-modules-extra kernel-network-modules" to omit those dracut modules from the run. I submitted a patch that was rejected, because apparently *some* places build an initramfs with 'dracut --no-kernel' and then append another cpio image that contains kernel modules, and the patch would break that. This is why we can't have nice things.)
On Sat, Jun 26, 2021 at 2:39 AM Valdis Klētnieks <valdis.kletnieks@vt.edu> wrote:
On Fri, 25 Jun 2021 16:05:45 -0400, Aruna Hewapathirane said:
Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs' Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules. Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
Does that make sense to you or anyone ?
For reasons understandable only to systemd, it tried to modprobe vmhgfs and didn't find it in the initramfs.
I think I am beginning to understand what went wrong, I installed VirtualBox and all my troubles started after this. Everything was smooth until I tried to compile a kernel. This is when things went sideways. So if your going to be compiling kernels stay away from VirtualBox on Debian. So as much as I do not care for systemd because it deviates from the original philosophy of do one thing but do that one thing really really well and better than anyone else policy to do everything under the sun ( and so far seems to do everything under the sun really well). It was 'me' who did this 'not' systemd. I am human, humans err ? Systemd is impressive and scares me at times but then I guess it comes down to personal choice ? I keep asking myself 'if' something bad happens to systemd, what are our options ? Think about it ? It is not like tweaking init using text files and echo eh ?
Does this configuration actually *need* a vmware shared folder in order to boot? Uh-uh no it does not.
I guess it is looking for that 'vmhgfs' because I installed VirtualBox. And something really interesting, I deleted VirtualBox or so I thought. I recompiled the kernel and checked, same 'bleeping' error. So reinstall VirtualBox and like some Voo Doo magic all my operating systems showed up which tells me Virtual Box keeps its folder and settings in your home folder. One lives and learns :-)
If so, you'll need to tell your dracut/ mkinitramfs/whatever command to include it. If not, you'll need to tell it to not look for it.
oh-kay so here are the latest developments. I told initramfs to load crc32c, told make menuconfig to disable Virtualization then recompiled and re-test and my kernel drops to a initramfs shell ( some progress yay ) then I ask initramfs to load 'vmhgfs' and it complains about 'can't find module not in blah-blah' and reverts to the original error, so I am back to square one. sudo vim /etc/initramfs-tools/modules ( in Debian ) for anyone curious :-)
(Particularly annoying - dracut will include that systemd unit even if you have a kernel that doesn't need any modules to boot because all the requisites are built-in, you said --no-kernel to not include kernel modules in the initramfs, and you gave it -o "kernel-modules kernel-modules-extra kernel-network-modules" to omit those dracut modules from the run. I submitted a patch that was rejected, because apparently *some* places build an initramfs with 'dracut --no-kernel' and then append another cpio image that contains kernel modules, and the patch would break that. This is why we can't have nice things.)
Valdis, if I apply that patch to 'my' kernel and it works then problem solved right ? What do you think ? And you do have nice things my friend, remember a young lady some years back wrote in her facebook... "Drove a Jag, did not crash" ? :-) I am going to keep working on this as and when I have the time I want to know how to fix this. We know it is because I installed VirtualBox, now I gotta find a fix :-( Thanks everybody - Aruna
Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs' Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules. Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
Does that make sense to you or anyone ?
For reasons understandable only to systemd, it tried to modprobe vmhgfs and didn't find it in the initramfs.
What Valdis says: I would expect that you can find a 'vmhgfs.ko' file somewhere in /lib/modules that systemd is trying to load, but can't because it wasn't included in the initramfs. *Goes looking for 'vmhgfs.ko' in package manager* Oh, so the vmhgfs.ko module is installed with open-vm-tools, not virtualbox-modules. TIL. Also it appears that vmhgfs is for VMware, not VirtualBox.
Does this configuration actually *need* a vmware shared folder in order to boot? Uh-uh no it does not.
I guess it is looking for that 'vmhgfs' because I installed VirtualBox. And something really interesting, I deleted VirtualBox or so I thought. I recompiled the kernel and checked, same 'bleeping' error. So reinstall VirtualBox and like some Voo Doo magic all my operating systems showed up which tells me Virtual Box keeps its folder and settings in your home folder. One lives and learns :-)
Just so we're on the same page: This module is for VMware, not VirtualBox. So I wouldn't expect (un)installing VirtualBox would affect this module.
If so, you'll need to tell your dracut/ mkinitramfs/whatever command to include it. If not, you'll need to tell it to not look for it.
Exactly. I don't know if/how you do this with systemd, but if you don't need the module at boot, systemd shouldn't be trying to load it.
I am going to keep working on this as and when I have the time I want to know how to fix this. We know it is because I installed VirtualBox, now I gotta find a fix :-(
Silly idea maybe (because I'm no systemd expert): In OpenRC you can blacklist modules in /etc/modprobe.d. Maybe something like this [0]: echo "blacklist vmhgfs" >> /etc/modprobe.d/vmware Looks like you can even add it as a boot parameter. And I would assume that it works with systemd as well, because Arch Linux uses systemd by default. [0] https://wiki.archlinux.org/title/Kernel_modules#Blacklisting Best of luck! -- Thomas
participants (4)
-
Aruna Hewapathirane -
Cindy Sue Causey -
Thomas Bracht Laumann Jespersen -
Valdis Klētnieks