This verkakte thing. It seems having this package installed and enabled
will break any eoan install. At least it broke all of mine, but in a way
that's caused great fun for the better part of a week now.

So, say you have prelink already installed. It's been there for years, never 
caused trouble, you forget it's even there. You then upgrade disco to eoan. 
Upgrade process seems to have gone quite smoothly - great. Getting on with your 
day.
Truth is you have a ticking time bomb.

Within 24h, cron.daily will have invoked prelinking of your new lib
versions. This will cause ~every useful binary on your system to
segfault for no immediately obvious reason. Syslog will tell you things
like this, which if you know the culprit is prelink makes total sense,
but if not adds to the mayhem:

Nov  7 01:11:14 plopp kernel: [  379.248086] ldd[5762]: segfault at 0 ip 
00007f2adace11f6 sp 00007ffecdb69978 error 6 in 
libc-2.30.so[7f2adace1000+178000]
Nov  7 01:11:14 plopp kernel: [  379.248094] Code: fe ff ff 90 f3 0f 1e fa 68 
19 00 00 00 f2 e9 21 fe ff ff 90 f3 0f 1e fa 68 18 00 00 00 f2 e9 11 fe ff ff 
90 f3 0f 1e fa 68 05 <00> 00 00 f2 e9 01 fe ff ff 90 f3 0f 1e fa 68 17 00 00 00 
f2 e9 f1
Nov  7 01:11:14 plopp kernel: [  379.248421] awk[5763]: segfault at 0 ip 
00007f957042b1f6 sp 00007ffc30077d58 error 6 in 
libc-2.30.so[7f957042b000+178000]
Nov  7 01:11:14 plopp kernel: [  379.248428] Code: fe ff ff 90 f3 0f 1e fa 68 
19 00 00 00 f2 e9 21 fe ff ff 90 f3 0f 1e fa 68 18 00 00 00 f2 e9 11 fe ff ff 
90 f3 0f 1e fa 68 05 <00> 00 00 f2 e9 01 fe ff ff 90 f3 0f 1e fa 68 17 00 00 00 
f2 e9 f1
Nov  7 01:11:14 plopp kernel: [  379.252793] apport[5764]: segfault at 0 ip 
000000323b007506 sp 00007ffe79f16ee8 error 6 in 
libpthread-2.30.so[323b007000+11000]
Nov  7 01:11:14 plopp kernel: [  379.252801] Code: fb ff ff 90 f3 0f 1e fa 68 
4b 00 00 00 f2 e9 31 fb ff ff 90 f3 0f 1e fa 68 4c 00 00 00 f2 e9 21 fb ff ff 
90 f3 0f 1e fa 68 4d <00> 00 00 f2 e9 11 fb ff ff 90 f3 0f 1e fa 68 4e 00 00 00 
f2 e9 01
Nov  7 01:11:14 plopp kernel: [  379.252822] Process 5764(apport) has 
RLIMIT_CORE set to 1
Nov  7 01:11:14 plopp kernel: [  379.252822] Aborting core

You go WTF, and figure a reboot seems the way to go here. Wrong, cause
now your system won't boot anymore. With Plymouth, all you see is an
apparent freeze. With recovery mode and console, you see a kernel panic.

As pointed out by OP, getting rid of prelink and its cache will sort out the 
issue. But before you can get to that point you have to make the connection, 
which might be immediate if you just installed the package, but far less 
obvious in the case of an upgrade.
And then, of course, start the recovery process of every affected install.

The prelink package ships defused, so it doesn't really do anything, but
once you activate its one purpose you have effectively set a timer to
brick your system. I'm not sufficiently familiar with the bug escalation
process in Ubuntu, but this to me sounds like release critical material
or similar.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1847611

Title:
  Installing prelink on a fresh eoan install gets me a "Core dumped"
  error

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/prelink/+bug/1847611/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to