On Fri, Apr 17, 2020 at 09:13:46PM +0200, Frank Krüger wrote:
Am 17.04.20 um 21:06 schrieb Michal Kubecek:
On Fri, Apr 17, 2020 at 08:27:38PM +0200, Frank Krüger wrote:
Am 17.04.20 um 20:08 schrieb sam@fnet.cx:
Hi,
I am getting this error message when I am updating a Leap 15.2 Beta system: "Error: Subprocess failed. Error: RPM failed: libgcrypt selftest: binary (0): Selftest failed (/usr/lib64/.libgcrypt.so.20.hmac)"
If I press "r" for retry the error just repeats over and over again, and if I press "i" for ignore all the successive packages zypper tries to install throw an error like:
"Abort, retry, ignore? [a/r/i] (a): i ( 99/331) Installing: xembedsniproxy-5.18.4.1-lp152.1.1.x86_64 ................................................................[error] Installation of xembedsniproxy-5.18.4.1-lp152.1.1.x86_64 failed: Error: Subprocess failed. Error: RPM failed: libgcrypt selftest: binary (0): Selftest failed (/usr/lib64/.libgcrypt.so.20.hmac) Ohhhh jeeee: ... this is a bug (global.c:150:global_init)
This is the second time this has happened. I had to reinstall the entire system because when I booted it up there was a kernel panic after this libgcrypt error. I thought it was something I had done with adding an incorrect repo but apparently not, because this is a clean install now. I now can't launch applications and can only use applications that were still loaded.
Known issues: https://bugzilla.suse.com/show_bug.cgi?id=1169569) related to libgcrypt and the regression of the SLE15-SP2/Leap 15.2 kernel, which has been hitting more and more people. As for the latter, I wonder that today's build 629.2 was published, even though it still provides kernel-default-5.3.18-lp152.10.4.x86_64 without a fix. By the way, KOTD for Leap 15.2 works fine.
I'm aware of the libgcrypt issue; do you have more information about the kernel regression?
Not a crash, AFAICS.
https://bugzilla.suse.com/show_bug.cgi?id=1168999, as well as the comment https://bugzilla.suse.com/show_bug.cgi?id=1163197#c73.
This might be but it only affects specific hardware. To be honest, I find it way more likely the crash reported here was caused by the fact that we use systemd as init and systemd is also linked against libgcrypt so that PID 1 process aborts which results in kernel panic - but that is not a kernel bug, it's expected behaviour. Michal Kubecek -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org