• Content Count

  • Joined

  1. The "proper" solution, i.e. one which requires neither downgrading the kernel nor masking new capabilities, is pretty straightforward - recompile Buster systemd packages against Linux-5.8+ kernel headers. I thought my original explanation ("systemd built against kernel headers older than 5.8 does not work") made this clear, in hindsight I should have been more explicit.
  2. Unfortunately upgrading systemd to the version from buster-backports doesn't help, that version has apparently still been built using older kernel headers: # systemctl --version systemd 245 (245.6-1~bpo10+1) +PAM +AUDIT +SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD +IDN2 -IDN +PCRE2 default-hierarchy=hybrid # systemctl show armbian-hardware-monitor.service >/dev/null Failed to parse bus message: Invalid argument # echo $? 1 (this is from a system which has been restarted several times since t
  3. Hello, In case you haven't heard about this yet, the fact that Armbian Buster Current comes with a 5.8 kernel means it suffers from this Debian bug: which us caused by combining this kernel version (as well as future ones) with systemd built using older kernel headers. This bug is mostly harmless but it does at the very least prevent Ansible from managing services on affected systems due to 'systemctl show' exiting with a non-zero exit code. Unfortunately with the official Debian Buster ker
  4. I was going to say that I had - I did update the u-boot package - but then I realised that the version number I saw on the serial console was not the same as that in the package description. Long story short, I forgot to actually write the u-boot image to the system disc on ALL the machines I have just switched to current... Interesting that it's only been the lime2 which failed to boot under those circumstances. Anyway, one quick dd session later and everything looks in order again. Thanks for having pointed me in the right direction!
  5. Hello, Earlier on this week I decided to upgrade the kernel and the DTBs on my Armbian cluster (quite a few different boards but all of them based on Allwinner A20) from buster-next-5.92 to buster-current-20.08.2. Having successfully completed the upgrade on several other servers, I moved on to the lime2 one - which promptly failed to come back up after the reboot. Connecting the serial console shows u-boot doing its thing without issues and the kernel beginning to boot, producing normal messages - only to suddenly freeze. When exactly the freeze occurs varies, sometimes it even ha