Hi, on my desktop server pc I was updating an during the process appeared a message that / partition was full, so I aborted the update but the partition resulte full, I deleted all files inside /tmp and rebooted, now it booted, grub ok, launched opensuse 15.6, I have a black screen, the server works becouse I can access the files from another pc by fstp.
what I have to do to or to delete to have it working again in graphic mode??
Deleting files from /tmp/ usually won’t gain much space. You may have retained too many snapshots and need to delete some to free considerable space.
Another place where excess can accumulate is in /var/log/journal/*/. How much space to be consumed there is controllable via /etc/systemd/journald.conf.d/.
If updates packages are being retained on the system, they must be cleaned periodically using sudo zypper clean
.
There’s a text mode utility ncdu designed for readily navigating through your filesystem to see where space is most used. Using it you may find runaway space consumption.
See SDB:BTRFS - openSUSE Wiki for checking BTRFS filesystem freespace.
manythanks, tomorrow I hope I can access with ctrl>alt>F2 or in recovery mode or booting from an usb stick with leap 15.6 and I will check /var/log/journal/*/.
I will run also sudo zypper clean
, but if I boot from an usb stick I suppose I cannot run sudo zypper clean
, can you suggest me a way to access in order to run sudo zypper clean
?
thanks for the link but I’m on EXT4 filesystem, how can I check freespace there?
and also how can I expand the / partition?
also I have separated partitions for / and /home and /srv
From a rescue boot to USB, before mounting installed / filesystem, run e2fsck on it.
Zypper clean mainly just deletes files (and directories) in /var/cache/zypp/packages/. There may be no rpms there. First check /etc/zypp/repos.d/*repo for the values of keeppackages=. There shouldn’t be any rpms to delete or space to recover if all are keeppackages=0. This process may be a bad idea at this point due to your aborted update process. Freespace should better be increased by other methods, so that it isn’t necessary to re-download packages to finish the update, in case networking is fouled by the abort. If updating can be completed, then you can safely go ahead with this.
Checking freespace on EXT4 is done the old fashioned way using df command, e.g. df / for your filesystem of interest.
There might be unneeded backups in /var/adm/backup/rpmdb/ and/or unneeded /var/log/*.xz files.
You may be able to gain substantial / filesystem freespace by uninstalling *-32bit, *-doc, *-devel, ghc* and/or texlive* rpms or rpms for other software that you never use. You can limit software not yet installed to only required rpms (exclude recommendeds) by an edit to /etc/zypp/zypp.conf to include solver.onlyRequires = true.
Check /boot/ for kernel count. If you have more kernels there than /etc/zypp/zypp.conf’s multiversion.kernels = calls for, you may have a non-working purge kernels service that needs to be fixed, or to simply uninstall one or more kernel rpms in excess of latest.
Any existing /var/log/updateTestcase-* directories older than from before your most recent successful system update or upgrade are unneeded.
If separate /home/ and /srv/ filesystems were not part of originally installing the OS, you may need to temporarily unmount them and confirm those two directories are then empty. Booted to rescue USB and not chrooted into the installed system they should already be unmounted and empty.
Booted or chrooted into the installed system, use coredumpctl to determine whether stale core files are wasting space.
Howtos for expanding EXT4 filesystems and partitions abound.
I installed timeshift and got /root disk full messages soon after when it just started writing backups to my root directory instead of my external hard drive on it’s own. It starts with no notice as soon as you set a schedule parameter.
Killed timeshift process and deleted timeshift backup directory it had created. All is now ok.
Make sure that timeshift was not runnung.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.