You are not logged in.
Kind regards to all,
After updating the system
# pacman -Syu
and rebooting, I get stuck at
Loading Linux ...
Loading initial ramdisk ...
I've tried several times with a clean install, but it never gets past the same point. I have always done the installation following the instructions very carefully with the only variation of,
# pacstrap base-devel networkmanager nano
I would be very grateful if you could guide me to solve my problem.
Please note that I am an inexperienced user.
Regards,
Alberto
ADDITIONAL INFORMATION
IBM ThinkPad 32Bit
SSD 128Gb
RAM 2Gb
Four partitions: 512M boot, 25G /root, 1G swap, 100G home
Last edited by trujillo-gc (2022-03-18 23:39:45)
Offline
I have some things you can check:
What is the model of your laptop? https://www.thinkwiki.org/wiki/ThinkWiki, Does your CPU have SSE2, SSE?
What architecture are you installing? pentium4 is for SSE2, i686 for SSE. pacstrap/pacman try to guess
the architecture, but sometimes they make mistakes..
Try to enable debugging on boot (remove quiet from the grub boot kernel options, increase the loglevel).
You have enough RAM, so I rule out decompression errors in the ramdisk.
Do you need a special driver to access your hard disk or SSD? Check when booting on the ISO, what
drivers are loaded for storage, then also load them in MODULES in mkinitcpio.conf.
Offline
Thanks for answer quickly.
Looks like I have work to do; give me some time and I'll get back to you with the results.
Regards,
Alberto
Offline
What I understand from the information that I have been able to obtain on the web that you provide me, is that my CPU supports both architectures SSE2 and SSE.
https://www.thinkwiki.org/wiki/Intel_Pentium_M_(Banias)
Offline
I have also modified the grub as you suggested, increasing loglevel and removing quiet.
# nano /etc/default/grub
GRUB_CMDLINE_DEFAULT="loglevel=7"
But it has not resulted in any positive change.
I looked carefully when booting on the ISO, what
drivers are loaded for storage, but I can't see anything showing the drivers for the ssd card.
So I'm afraid I haven't made much progress.
Offline
I wonder if as part of doing the update, the system failed to build the initcpio's again for the new kernel. As far as I recall the magic runes to make that run again more manually are 'mkinitcpio -p linux'
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
I used to rock a boot partition mounted as read only, and as I recall the upgrade process would bail out, and force me to remount it as rw before it would complete. If you don't have it mounted, I suspect it would simply create a boot folder in the system root and fill it with files.
Edit: Hmm, yes, if you had poplated the root folder with a new kernel and initcpio, but left the contents of the boot partition as was, and then booted off that boot partition, would it have any difficulty in handing over to the real init?
Last edited by levi (2022-02-28 22:08:46)
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
Since upgrading linux and linux-pae to 5.16 just a black screen is presented after choosing to start these kernels from grub. Prior linux and linux-pae versions <5.16 are working fine.
For now I am working with linux-lts (5.15 series) and put it to ignore list in pacman.conf.
Any ideas?
Offline
I can see you guys have been busy during my absence. I apologize for not checking.
I must say that I am a bit lost with your discussion, remember that I do not have much experience
Offline
I'm sorry but work obligations keep me busy at the moment; I promise to update you as soon as I have some time to follow your suggestions.
Offline
I also asked my question in the general Archlinux forum, and they gave me this suggestion.
Last night I reinstalled grub like this...
# grub-install --target=i386-pc /dev/sda
# grub-mkconfig -o /boot/grub/grub.cfg
But nothing has changed.
Offline
Since upgrading linux and linux-pae to 5.16 just a black screen is presented after choosing to start these kernels from grub. Prior linux and linux-pae versions <5.16 are working fine.
Are you using linux and linux-pae on the same machine? Or across different machines. I note they're not marked in the package as conflicts, but I suspect strongly that they should.
FWIW, I'm currently rocking linux kernel 5.16.8 for a couple of weeks and it's working for me.
I also asked my question in the general Archlinux forum, and they gave me this suggestion.
Last night I reinstalled grub like this...
I'm not sure why you want to reinstall grub. Grub seemed to be working for you, and it was only after grub handed control over to the initcpio that it all falls flat.
As I wrote earlier, there is a way to rebuild your inicpio in case that's bad:
mkinitcpio -P
Last edited by levi (2022-03-01 21:08:25)
Architecture: pentium4, Testing repos: Yes, Hardware: EeePC 901+2GB RAM+OS half on the SD card.
Offline
It's all the same machine, an Acer Travelmate 290 with a Single Core Intel Pentium M (Centrino) CPU, Intel 82852/855GM Integrated Graphics and 2 GB RAM running archlinux32 pentium4 with i3wm. Things with linux kernel were fine, until update. Didn't test manually building initram as suggested, cause never had a problem since with pacman doing that job
Offline
Hello. I have this problem too on several Dell laptops-- one is a Pentium M, one is a Centrino and one is a Pentium II. They are D600, 600M and Latitude CPi models. The D600 and 600M have 1GB of ram each and the CPi has 512MB. All of these machines used to run ArchLinux without issue and then briefly took a trip through Fedora-land (FC14 to be specific) and then recently I have tried to reinstall them with latest ArchLinux32-2022.02.01 distro without success. The installation goes smoothly, seemingly without trouble until the post-install reboot. At this point, they all hang in the same way as the OP states-- after the initrd is presumably unpacked, the machine is locked up tight. CTRL-ALT-DEL won't reboot it, a quick press on the power button does nothing-- only a long press on the power button to completely power off is the way out.
I have also tried going back several ArchLinux32 releases including all the way back to 2017.05.01 which fails the same way.
Through this adventure I have also tried Debian 11, 32-bit and this will install and run without issue on all three machines. As also noted, Fedora 14, 32-bit also installs and runs on them. So, I think the hardware is good. There seems to be something unsavory with the initrd that is being loaded after install.
Chris
Offline
Just made a new ISO archlinux32-2022.03.02-i686.iso, hope it doesn't have the same trouble.. I'm currently testing.
What's quite curious to me is that both an ISO from 2022 and 2017 should behave in the same way, hard to be true. :-)
Currently testing on virtual machines, I hope it shows also there, otherwise I have to kick one of the really old and noisy machines... ;-)
Offline
Thank you @abaumann! I will give this a try later today. Have to at least pretend to be focusing on ${dayjob} for a few more hours. I too was surprised by the 2022 & 2017 similar behavior so I will also repeat that experiment. I could have been fooled. It is a rather time consuming test loop since need to do a full install before it can be tested-- but I will do, across all three machines and report back.
Can you comment on what you have changed?
Chris
Offline
Really strange. I was on real hardware (a pentium II without SSE2), doing a mkinitcpio in a chroot. No problems to reboot.
I really need some log messages to find out what's going wrong here..
Offline
yes-- agreed. I have a new issue trying with your 2022.03.02 build on the D600 laptop. Now it hangs during installation boot after "::Triggering uevents" and I get a couple of *ERROR* messages, one about "No UMS support in radeon module!" and "VGACON disables amdgpu kernel modesetting."
I have tried a number of kernel command line options such as "nomodeset" and "amdgpu.dc=0" with no improvement.
So, I am stuck even earlier in the process now and therefore can't provide any input on whether the ISO from yesterday allowed me to get through install and then boot.
I'll try some other hardware too.
Offline
The only similar thing I had lately was on the Alix board with AMD Geode GPU, were the kernel completely disabled reads to the Video BIOS, rendering Xorg unusable..
Nobody seem to give a damn it seems and test things on old hardware (aka "move fast and break things").
Offline
yes-- I am having bad luck here now with old hardware it seems. Can't get the installation to go on the Dell CPi either. This appears to be a not enough RAM issue. I was mistaken in my earlier post, stating this machine had 512MB. It only has 256MB so I think with a modern kernel, I am out of luck there. FC14 did install and run but that is a 2.4 kernel from Antiquity.
Thanks for your help in any case. I think I may have to devise a method to install to these laptop HDs on a more capable machine and then move the HDs back to the laptops for run-time. That's a bigger project that will take a while.
My ultimate goal was to get one of these old machines running a distro with flexible package support so that I could hopefully get some PCMCIA and MTD tools in place to access an old linear memory PCMCIA card (which each of these laptops have a slot for).
Perhaps some of this technology really does become too old :-(
Chris
Offline