Missing libcrypt.so; Reinstall?

[anw1652]

I installed a Debian Bullseye (stable) system and, after rebooting and doing some superficial checks, everything seemed to go fine. I've installed and administered many *nix systems (and most were Debian, but it's been a while) and this was my home system, so I didn't test very extensively. I left for a few weeks and when I returned, the system had rebooted but there was no login. After booting into rescue mode, the problem seemed to be that the system couldn't find libcrypt. I also noted that I had neglected to change the repos in apt's sources.list from "testing" to "stable" (I had intended to upgrade to Bookworm manually) and there's a cron job that does automatic updates. When in (Bullseye's) rescue mode (chroot'd to /target), there is a bunch of stuff in /usr/lib/x86.../, but no libcrypt.so. Also, /usr/share is linked to /usr/hd_mnt, and it is empty (hence, no man pages). So, after rooting around on the internet, I think I got caught in the usrmerge kerfuffle when one of the updates occurred and Bookworm had expected that to have happened.

At this point, I decided to just reinstall Bullseye, then do the manual upgrade to get to Bookworm. I thought I remembered Debian offering an option during the install to retain all the existing drive partitions and user data, but I don't see anything like that now. Therefore, my current idea is to run the Bullseye installation up to the step of partitioning the drives, skip that step, then continue with the rest of the installation.

So, my questions are:

1. Is there a simpler way to either restore the Bullseye install or go to the Bookworm distribution?
2. If not, are there any obscure (in particular, detrimental) ramifications to the above skip-step idea?
3. If not, at the end of the partitioning step, does the installer expect the disk partitions to be mounted, and, if so, where? Off / or /target or...?

[anw1652]

I installed a Debian Bullseye (stable) system and, after rebooting and doing some superficial checks, everything seemed to go fine. I've installed and administered many *nix systems (and most were Debian, but it's been a while) and this was my home system, so I didn't test very extensively. I left for a few weeks and when I returned, the system had rebooted but there was no login. After booting into rescue mode, the problem seemed to be that the system couldn't find libcrypt. I also noted that I had neglected to change the repos in apt's sources.list from "testing" to "stable" (I had intended to upgrade to Bookworm manually) and there's a cron job that does automatic updates. When in (Bullseye's) rescue mode (chroot'd to /target), there is a bunch of stuff in /usr/lib/x86.../, but no libcrypt.so. Also, /usr/share is linked to /usr/hd_mnt, and it is empty (hence, no man pages). So, after rooting around on the internet, I think I got caught in the usrmerge kerfuffle when one of the updates occurred and Bookworm had expected that to have happened.

At this point, I decided to just reinstall Bullseye, then do the manual upgrade to get to Bookworm. I thought I remembered Debian offering an option during the install to retain all the existing drive partitions and user data, but I don't see anything like that now. Therefore, my current idea is to run the Bullseye installation up to the step of partitioning the drives, skip that step, then continue with the rest of the installation.

So, my questions are:

1. Is there a simpler way to either restore the Bullseye install or go to the Bookworm distribution?
2. If not, are there any obscure (in particular, detrimental) ramifications to the above skip-step idea?
3. If not, at the end of the partitioning step, does the installer expect the disk partitions to be mounted, and, if so, where? Off / or /target or...?

[Aki]

Hello,

It seems you cannot boot your Debian Stable (Bullseye) installation.

In Debian there’s no automatic “restore” procedure for a broken installation: you have to do it manually, if the system is in a fixable state. The installation disk started in rescue mode is a tool suitable for this task, but it could be not easy and you have to know exacltly what you are doing.

This is the Debian package containing the libcrypt library: https://packages.debian.org/bullseye/libcrypt1

If you do not need to retrieve personal data from the broken installation, it is probably easier to reinstall Debian (overwriting previous installation), choosing the Debian testing release because it’s the one you are interested in it.

[anw1652]

Aki, thanks for the reply.

I would prefer to preserve the data, but I did back it up onto a partition that I can wait to mount until after the installation.

The disk has one large physical partition subdivided using LVM. When I go into rescue mode from the Bullseye DVD (literally a physical DVD), it shows this as I had set them up. However, when I use the Bookworm installation (which is on a USB memory stick prepared on a Windows machine with Rufus), it just shows the physical partitions, not the LVM subdivisions. Do you have any idea why that is? Does it have something to do with the USB booting with EFI?

[Aki]

The disk has one large physical partition subdivided using LVM. When I go into rescue mode from the Bullseye DVD (literally a physical DVD), it shows this as I had set them up. However, when I use the Bookworm installation (which is on a USB memory stick prepared on a Windows machine with Rufus), it just shows the physical partitions, not the LVM subdivisions. Do you have any idea why that is? Does it have something to do with the USB booting with EFI?

I suppose because the installer has a different behavior if is started in rescue mode or it is started in install mode.
I beg your perdon, but I have to remember you that it could be safer to backup your personal data in an external disk before the installation, of course.

[iv_kravchenko]

It's probably a bit to late. But put it here for the next generations.
The problem is fixable:

https://www.lisenet.com/2023/libcrypt-s ... -bullseye/