I’m curious why tumbleweed wants to update ALL of my installed packages… About 3000 of them. It looks like the last digit of the version number for every package changed by 1. But I’m sure every package didn’t actually get changed.
I’m just wondering why… And also making sure this is intended and not a problem on my computer.
Thanks.
As suggested elsewhere in another thread, it is psobably a change of compiler. And thus everything is build anew. And yes, that last number in the version is the build number.
yes gcc 9.1 is now being used for all packages
Thanks! (Sorry about asking a question that’s been covered… I did a quick search, but apparently not very well)
THIS IS MADNESS.
There’s an unresolved bug with kernel 5.0+ and GCC 9 that causes bcache to seriously corrupt data. Fedora found it and Arch Linux has pulled GCC 9 from testing as a precaution:
Somehow everyone with OpenSUSE missed this and went and compiled everything with GCC 9 anyway. Unless they’ve done something special (compiled everything except the kernel?) they’re going to release a bcachepocalypse upon Tumbleweed users.
And if they have done something special to avoid the bug, somebody needs to step up and make an official announcement. I’ve put a halt to updating every Tumbleweed machine I’m in charge of until there’s some statement that it is safe and I’ve posted a warning on Reddit. Even if they’ve done something special, shipping GCC 9 is a bad idea since someone might unknowingly compile their own kernel with it and kill their system (that’s what Arch was worried about).
I wish there was some big red panic button I could push somewhere. Is there any official way to contact the Tumbleweed people and ask them what the heck they’re doing with this update?
Unless this is a very recent install, you should be able to boot to the previous kernel. And that was not recompiled for this update.
Hi
There is the kernel mailing list…
https://en.opensuse.org/openSUSE:Communication_channels
https://lists.opensuse.org/opensuse-kernel/
1102247 – fstrim on bcache enabled file system produces an Oops (?)
“zypper dup” found “6622 packages to upgrade, 6 new, 1 to remove.” So I hesitated to update anyway.
On the other hand data corruption is limited to systems using SSD caching of HDDs. Will it be safe to install the snapshot on systems never using this feature?
I did the update on two computers, one with Plasma and one with Gnome both succeeded as alwaysrotfl!
I went ahead and upgraded one TW then running still on 5.0.13 to 20190529. It only upgraded about 800 packages. Most icons are missing from its IceWM panel. The only recognizable ones are the starter and mail.
Don’t worry. In any case, it was in the German language section, thus I am not surprised you did not find it lol!
I got hit by the bcache apocalypse. My /home xfs is now corrupted. What do I do now?
If web searching isn’t up to the task of finding a solution to XFS corruption, start a thread with an appropriate subject for attracting those who might be interested and able to help with an XFS corruption problem. This one is not.
I’ve updated a TW/KDE over the past 24 hrs and haven’t seen a problem… yet.
Maybe some kind of b-cache problem might affect some systems, but I highly doubt a gcc problem should affect too many systems…
Most rpm packages should contain pre-compiled code (so installing mainly just writes and over-writes files).
That’s what’s supposed to be happening on OBS… packaging but also typically compiling.
Any packages which might require a compile on the fly could be affected, but I don’t see very many of these.
People should know that a compiler like gcc mostly sits on a machine doing nothing… It’s used once and only once to compile code, but once that code has been compiled the compiler (eg gcc) isn’t used again.
TSU
The corruption of my filesystem is a Problem, but it is not the biggest one I have.
The Problem I (and maybe everyone using bcache) have is, whenever I start current tumbleweed my Files get corrupted again and I (we) will probably loose data.
I think most of you really underestimate this Problem! !DATA LOSS!
Is there any way I can solve that Problem?
I don’t doubt that it is a serious problem.
I’m not using “bcache” here. I guess that’s my basic conservatism (or Luddite tendency, or “if it ain’t broke, don’t fix it”).