May 2021 - openSUSE Users - openSUSE Mailing Lists

Compositing crashing KWin on Login
by Matthias Bach 29 May '21

29 May '21

Hi all, For a while now I have this weird issue where, when logging into KDE after system boot, KDE is in a weirdly broken state: compositing is disabled (no wobbly windows, only half my windows are restored (curiously, Firefox always is, Konsole is always missing), and the desktop itself is not being drawn, resulting in remnants of moved windows staying on it. This happens when using the official NVIDIA driver from the openSUSE-NVIDIA repository on a GTX 1060. I have been using this same set-up for years but the issue only popped up a couple months ago without me knowingly changing anything. It also survived the move from Leap 15.2 to Leap 15.3 Beta / RC. Only my main user seems to be affected by this. Two test users I use irregularly don't seem to show the same behaviour. When going to the compositing settings in that state, I get a message that compositing crashed KWin and OpenGL detection has thus been disabled, see the German-language screenshot attached. Reactivating the detection and either doing a logout / login or running "kwin_x11 --replace" will result in compositing working again. I also have a slight impression that this is more likely to happen when there is disk activity while initiating the login. But this might also be a red herring. Does anybody have any hints what log files might help to home in on the issue causing this? I have a suspicion that I have some non-default configuration, that I am not aware of, that either directly causes this or exposes some underlying bug that is otherwise not triggered. Kind Regards, Matthias -- Dr. Matthias Bach www.marix.org „Der einzige Weg, die Grenzen des Möglichen zu finden, ist ein klein wenig über diese hinaus in das Unmögliche vorzustoßen.“ - Arthur C. Clarke

1 1

Re: Moving from Freenode to Libera.Chat
by Carlos E. R. 27 May '21

27 May '21

On 27/05/2021 13.37, Richard Brown wrote: > On Thu, 2021-05-27 at 13:31 +0200, Per Jessen wrote: >> Carlos E. R. wrote: >> >>> Again. Pidgin requires the password BEFORE a connection is >>> established. >> >> If nothing else, this is going off-topic here - FWIW, it worked just >> fine with konversation. No userid, no password, but there was a >> third >> step to verify my email address. > > Indeed - please can we relocate the "Carlos doesn't know how to connect > to IRC" portion of this thread to the Support list please? As long as somebody helps me there... Connection doesn't work: https://susepaste.org/81396729 -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)

3 5

We knew there was something funny about the new Tbirds and gpg ...
by David C. Rankin 27 May '21

27 May '21

Icarus moment: Mozilla Thunderbird was saving OpenPGP keys in plaintext after encryption snafu Cockup has since been patched in latest release 6 Comments https://go.reg.cx/tdml/dfd67/60d51c7f/70bba22e/3ZNf Gives me the shivers about ever updating.... -- David C. Rankin, J.D.,P.E.

3 8

IRC moving from freenode to libera
by Per Jessen 27 May '21

27 May '21

... continued from factory. Carlos, you have the wrong server specified. It is 'irc.libera.chat', not 'libera.chat'. -- Per Jessen, Zürich (14.6°C) Member, openSUSE Heroes.

4 4

bind mount and yast partitioner
by Marc Chamberlin 27 May '21

27 May '21

Hello, I am running a number of OpenSuSE systems on my SOHO network at home and I keep hitting a small but annoying problem that I wonder if there is a better workaround/solution to. This occurs on all versions of OpenSuSE - 15.0, 15.1, and 15.2. I am not running 15.3 yet. All systems are x64 architecture. I want to access the complete file systems of each computer via samba and autofs, which means I want to be able to access the "root" directory / on each system. Some years ago Samba stopped me from designating / as a samba shared directory and exporting it. (I don't understand why except from Googling it appears someone decided that to allow /, to be exported as a Samba share, was a security problem. I get so annoyed at all the layers of security and obfuscation so have not tried to track down what is exactly going on with Samba and the "root" / directory. Please don't try to dissuade me and tell me about the security risks of making the "root" / directory available for mounting, this is a SOHO network and I solely control all the computers on it.) I know Samba will allow me to define a share for the "root" / directory but it doesn't seem to work, at least autofs will not recognize/mount it and still reports the directory does not exist if I try to cd into it. So to work around this I do a bind mount of / as /slash and samba then allows me to export /slash as a samba share. In my various configuration files I use the name "slash" as the name of the exported Samba share and I create a directory /slash so that my configuration files will work regardless of whether /slash is a bind mount or not. At least that is what I think I am doing. The config line in fstab looks like this - / /slash none bind smb.conf has the following definitions for the root directory - [slash] available = Yes browseable = Yes comment = Root directory inherit acls = Yes path = / read only = No create mask = 0755 force user = root autofs configuration for mounting the root directory / typically looks like the following, only names of the remote host changes. slash -fstype=cifs,rw,dir_mode=0777,file_mode=0777,credentials=/etc/smb1.auth ://bigbang/slash This works fine for autofs and most tools that work on the host or across network file systems. All except, that is, the YaST2 partitioner. For some unknown reason it picks up the bind mount /slash and automagically reassigns all my mount points to be under /slash. So for example if I tell it to mount my home partition at /home the partitioner will reassign my mount point to be at /slash/home. More importantly, the partitioner reassigns / to be mounted as /slash. Then if I want to make any further change to partitions and save those changes, the partitioner will complain that the mount point for / is undefined and it wants to force me to define a / mount. The only workaround I have found is to remove the bind mount from fstab and then reboot. After reboot I can then use the YaST partitioner to make the changes I want, since because without the bind mount the partitioner now sees all the mount points of different partitions as subdirectories under / and not under /slash/. But without the bind mount defined in fstab, tools like autofs will no longer mount the "root" / directory though I do not grok why. So again I have to add back in the bind mount to fstab so that Samba shares will again work, and again I have to reboot. I have tried restarting all the appropriate services individually, such as nmb, smb, autofs, and network but no joy getting the the remote host's "root" / dir to remount, so I resort to rebooting the systems (a PITA) and that does the trick... Perhaps I am missing some magic trick or there is a better way to accomplish what I want to do, which ideally is to configure Samba to really share the "root" / directory. Sorry this is complicated and will be difficult to grok, it is for me! Any help would be much appreciated, I would love to once and for all resolve this irritant because I have to rediscover these issues and their resolutions each time something like autofs breaks or I want to use the partitioner. Thanks, Marc

6 13

php7 zmq (ZeroMQ) for LEAP 15.2, anyone?
by Adam Tauno Williams 26 May '21

26 May '21

Anyone know where I can find a php7-zmq package for LEAP 15.2? I'd struck out in the usual places and it doesn't want to build from PECL/phpize. This is the last dependency to forward port an application of an old server. :( -- Adam Tauno Williams <mailto:awilliam@whitemice.org> GPG D95ED383 Systems Administrator, Python Developer, LPI / NCLA

1 1

email of bugzilla of opensuse
by Stakanov 26 May '21

26 May '21

I am receiving my bugzilla reply mail on a historical freenet account that recently does not work well anymore (freenet causes me dns problems, no idea why). I tried to change the email to this one (disroot) but no luck up to now, even in my account settings no way to change the original historical email to something valid and new. Does anybody know a procedure to have this done? I have seen not further indication.

2 7

Why can't the process "sync" be killed?
by Carlos E. R. 25 May '21

25 May '21

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 It is just a fact, it can not. Situation. The system is running the process "texpire", which is part of leafnode package, an nntp proxy server. texpire simply looks at every posts and deletes those that are old according to certain rules. This is very intensive on the disk, looking at 1.2 million files. It runs for half an hour. If during that process I issue a "sync", it does not succeed till texpire ends. It can not be killed. Telcontar:~ # time sync ^C^C^C On another terminal: Telcontar:~ # killall sync Telcontar:~ # killall sync Telcontar:~ # killall -9 sync Telcontar:~ # killall -9 sync Telcontar:~ # It has to run till completion. - -- Cheers Carlos E. R. (from 15.2 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCYHN4lRwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfV1sMAn2cBZmEwd/2LPAIbwlZA xOYqhSPeAJ9JEDEbDbVTBxs9Ec7oK20KZpqWkQ== =Rcb3 -----END PGP SIGNATURE-----

7 16

named - writeable file 'master/2001.db8.1234.1': already in use: /etc/named.d/enidan-reverse.conf:40
by Per Jessen 24 May '21

24 May '21

I run a named with multiple views (four) - in such a setup, you cannot have independent / out-of-view zones so each view includes common zones, such as 'enidan-reverse' (just reverse zones). For the reverse ipv6 zone, we want to do dynamic updates - hence we have 'allow-update { localhost; };' for that zone. Previously, named or named-checkconf did not complain, but I am attempting an upgrade to 15.2 and the new named-checkconf complains. Plenty of hits with google, but no solutions other than disabling the checkconf. Some people suggest adding 'DISABLE_ZONE_CHECKING=yes' to /etc/sysconfig/named, but that variable is not supported in our named setup. For now, I need the named update more than the dynamic ipv6 updates, so I'll probably just disable it, but surely there is a way out of this? -- Per Jessen, Zürich (16.6°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland.

2 2

Automatically install debuginfo/debugsource packages?
by Niklas Haas 24 May '21

24 May '21

Hello, I do a lot of debugging, and I generally find it annoying to have to wait for debuginfo to become available. I'm used to distros such as Gentoo where I can globally enable debuginfo for all packages, and never have this problem ever again. I know that in recent times, openSuSE has started shipping gdb with the annoying `libdebuginfod` which attempts solving this problem but fails spectacularly. As such, I don't consider it a solution, for a number of reasons that are besides the point. (To anybody finding this email via google: comment out /etc/profile.d/debuginfod.sh to disable it) What I _want_ to have happen is for `zypper in <package>` to automatically pull in `<package>-debuginfo` where available. Is there any way at all to accomplish this? Thanks, Niklas P.s.: It's possible with some parsing of the `zypper packages -i` output to construct a command line that installs `<package>-debuginfo` for all installed packages. However, this still has annoying shortcomings: 1. Uninstalling packages doesn't auto-clean the (no longer needed) debuginfo packages, thus causing them to continue littering the system over time. 2. This has to be re-run periodically in order to keep up with newly installed packages.

4 4