Tue, 27 Jun 2023 20:35:55 +0000 (UTC) Robert Webb via openSUSE Users <users@lists.opensuse.org> :
On Tue, 27 Jun 2023 15:11:06 -0400, bent fender <ksusup@trixtar.org> wrote:
[...] then with both packman disabled:
# zypper dup [...] Illegal instruction (core dumped)
THEN, both enabled and a reboot: https://paste.opensuse.org/pastes/658a2cde8111
Does this speak of memory problems (RAM, not me)?
Don't know. When I had bad RAM, it appeared as BTRFS filesystem errors during 'zypper dup'.
THEN # zypper dup [...] Nothing to do.
How can I get load of failed scriplets (in the susepaste) followed by 'nothing to do'?
Did you notice that the last line of your zypper dup paste, after all the nasty errors, was Zypper saying like, "You just need to reboot to finish this up. Yeah, everything is fine."? But, it probably exited with an error code (invisible).
I use this to report any error: zypper dup || echo "($?)"
Those scriptlet errors don't typically leave any state info behind that zypper knows about, so the next time it runs, it's as if the previous run was successful. So, you have to notice the errors when zypper runs, but of course they have scrolled off the screen.
Your susepaste showed a core dump followed by "Failed to start transient service unit: ...". Maybe I can assume that a post-install scriptlet (of xdg-desktop-portal?) was trying to start a service. I just wonder in general about the zypper update process; Do some errors happen because current processes are running that are old versions of software that has been updated, and so fail sometimes when interacting with the new stuff? And so after a reboot, everything is actually fine? -- Robert Webb
Thanks for the details, mostly over my head but I have noticed increasing segfaults over the past couple of months and not just while driving zypper nor only with Leap/Tumbleweed