Hey Geekos,
First of all, sorry for this cross-post, but it is relevant to all of the
lists specified.
I am just wondering if my experience is so much different, or if I use the
distro differently than other users, that it seems so much stable for me.
Often i see people recommending against "zypper dup", recommending
"tumbleweed-cli" instead of "zypper dup", using "opi" for codecs install
instead of 1-click install on opensuse-community, and so on, because the
system "could" or "will for sure" break at some time. Why there is a need for
all this fancy and new stuff and is it possible that these combinations and
diversity increases possibility for issues?
Given all of these, I should have had my home machines broken already a
million times and burned to death.
- I am using Tumbleweed
- I have all these repositories enabled: [1] (yes, no priorities currently)
- I have "zypper dup", "zypper inr", "zypper ve" in my daily cron job
- I have enabled vendor changes in zypper conf
- I have disabled multiversion.kernel in zypper conf (keeps kernel packages
with their dependencies and/or DKMS clean and working)
- I have disabled snapper as a whole
- I have installed some VMWare software from their crappy .run packages, which
means some parts are compiled from source
During my over-15 years-long experience with *SUSE, i never experienced an
unbootable system or a serious issue, except:
- Nvidia driver issues before they were packaged for *SUSE and you had to use
the .run installation ( ancient SaX2 times, before Xorg autoconfiguration ). On
my gaming system i have just enabled nvidia repos and forgot about it.
- Akonadi issues because of my experiments with bleeding edge mariadb in
combination with my mailbox containing few million e-mails, also many years
back
- Btrfs issues (total system crash) when it was unstable and i had enabled
snapper an ran out of space - which was resolved by reinstall and re-mounting
my /home/ (i am not using snapper since then)
So my question is - have the RPM dependencies gone weaker than they were
"before" or something? Could that combination of all the emergency safe
features, like btrfs snapshots in combination with multiversion kernel and/or
packages, tumbleweed-cli, opi, etc...? Why are RPM deps not enough to keep the
system solid anymore?
My experience with *SUSE has always been rock-solid-stable and one of the very
few things i do manually from time to time is firing up yast2 sw_single,
removing unmaintained packages and checking change-logs of some packages i am
interested/care of.
Take this as a congrats and thanks for all the maintainers, developers and
everyone taking care. And most importantly, the OBS, which does pretty neat
job with automatic package dependencies and all that stuff. One more time -
thank you OBS and all the people developing and working on it! :)
And while i am doing this braindump, please revive SUSE Studio :P :)
And last, but not least - thanks Packman people :)
[1]: https://paste.opensuse.org/view/raw/38599835
Regards,
Gryffus
