[SOLVED] Encrypted luks2 partition won't boot

[Danielsan]

Hi guys,

I have a very annoying issue, I use a Debian VM to handle some specific task. It is Sid on Virtualbox with macOS as host. Today after a reboot it stopped to boot because after inserting the password I received this warning:

Code: Select all

libgcc_s.so.1 must be installed for pthread_cancel to work
aborted

And It requested again the password.

I chrooted and reinstalled a bunch of packages that have that library, I also installed GNULIB that should handle pthread tasks, and did a fsck; but none of those worked and I don't understand how to fix it.

Any ideas?

Everything is on / with a very small swap partition.

[p.H]

Maybe sid is broken again ?

libgcc_s.so.1 seems to be a "soft" dependency of libpthread (not shown by ldd, loaded only when needed). Did you check if it is included in the initramfs ?

[Danielsan]

Hi,

thanks for your reply.

Eventually came a bug report:

https://bugs.debian.org/cgi-bin/bugrepo ... bug=950551

And SID got broken again because a missing hook in the initram image.

I am bit of disoriented about this behavior, however I was able to run again my box but I am not sure if I will able to re-run it again...

Other reports:

https://salsa.debian.org/kernel-team/in ... 5ef9ccc015

[p.H]

Eventually came a bug report:

A change in a package broke another package. That's what sid/unstable is for.

I am bit of disoriented about this behavior, however I was able to run again my box but I am not sure if I will able to re-run it again...

What do you mean ? According to the links you posted, the issue has been fixed.

[Danielsan]

A change in a package broke another package. That's what sid/unstable is for.

I am agreed however it doesn't seem very smart throwing a change and see what happens without caring anything else.

What do you mean ? According to the links you posted, the issue has been fixed.

Everything works properly however none all the suggestions worked for me, hence I was able to reload Debian but without understanding which fixing actually worked.

[p.H]

The fix was simply to update the broken package.

[Danielsan]

From my understanding the fix was to go on an older kernel creating a symlink of that library where missing and regenareting the initramfs for the broken kernel...

[p.H]

The kernel was not broken. The initramfs was. Booting with an older kernel (using an older initramfs) was just a temporary workaround in order to boot the system. The real fix was to update the package which needed to be updated.