Closed art-lucas closed 3 months ago
art-lucas
commented
3 months ago After some browsing, I see that /dev/dm-X entries are typically LVM devices. If that's the case (?), why does my system have them considering that I am not using LVM (see lsblk above)?
art-lucas
commented
3 months ago I know that PopOS creates a LVM group (data) under the 3rd partition, here nvme0n1p3 with only one logival volume (root), which I found useless and luks-formatted nvme0n1p3 directly. Could that be the reason?
art-lucas
commented
3 months ago After killing the above timeshift process and let it run for a while:
# ps ax | grep timeshift
8377 ? DN 0:00 mount /dev/dm-0 /run/timeshift/8353/backup
34251 pts/0 DN 0:00 mount /dev/dm-0 /run/timeshift/34227/backup
34295 pts/0 DN 0:00 mount /dev/dm-0 /run/timeshift/34271/backup
36828 ? SNs 0:00 timeshift --check --scripted
36853 ? DN 0:00 mount /dev/dm-0 /run/timeshift/36828/backup# ll /run/timeshift/
drwxr-xr-x 2 root root 40 Mar 31 13:31 14761/
drwxr-xr-x 2 root root 40 Mar 31 14:00 18628/
drwxr-xr-x 2 root root 40 Mar 31 15:00 26799/
drwxr-xr-x 2 root root 40 Mar 31 15:43 34160/
drwxr-xr-x 3 root root 60 Mar 31 15:43 34227/
drwxr-xr-x 3 root root 60 Mar 31 15:43 34271/
drwxr-xr-x 3 root root 60 Mar 31 16:00 36828/
drwxr-xr-x 2 root root 40 Mar 31 17:00 43132/
drwxr-xr-x 3 root root 60 Mar 31 13:00 8353/
drwxr-xr-x 2 root root 40 Mar 31 13:00 8490/
drwxr-xr-x 2 root root 40 Mar 31 13:01 8829/
drwxr-xr-x 2 root root 40 Mar 31 13:02 9086/# ll /run/timeshift/*
/run/timeshift/14761:
total 0
/run/timeshift/18628:
total 0
/run/timeshift/26799:
total 0
/run/timeshift/34160:
total 0
/run/timeshift/34227:
total 0
drwxr-xr-x 2 root root 40 Mar 31 15:43 backup/
/run/timeshift/34271:
total 0
drwxr-xr-x 2 root root 40 Mar 31 15:43 backup/
/run/timeshift/36828:
total 0
drwxr-xr-x 2 root root 40 Mar 31 16:00 backup/
/run/timeshift/43132:
total 0
/run/timeshift/8353:
total 0
drwxr-xr-x 2 root root 40 Mar 31 13:00 backup/
/run/timeshift/8490:
total 0
/run/timeshift/8829:
total 0
/run/timeshift/9086:
total 0I ended up reinstalling the OS with the standard PopOS layout, i.e., with LVM group/lv @ /dev/data/root instead of luks-formatting nvme0n1p3 out of the box:
# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
...
...
nvme0n1 259:2 0 3.6T 0 disk
├─nvme0n1p1 259:3 0 1022M 0 part /boot/efi
├─nvme0n1p2 259:4 0 4G 0 part /recovery
├─nvme0n1p3 259:5 0 3.6T 0 part
│ └─cryptroot 252:0 0 3.6T 0 crypt
│ └─data-root 252:1 0 3.6T 0 lvm /home/linux/VMWare
│ /home
│ /
└─nvme0n1p4 259:6 0 4G 0 part
└─cryptswap 252:2 0 4G 0 crypt [SWAP]So far, after extensively using btrfs & timeshift, I have not seen the issue yet.
timeshiftmint22 24.01.1, latest to date/usr/bin/timeshift-autosnap-aptinstalled and configuredbtrfs-progsv5.16.2LUKS2.4.3The following issue happened before reinstalling the system from scratch -- for a different reason. I thought that perhaps the filesystem was corrupt or there was a misconfig due to the fact that I got rid of LVM in the previous install, so I didn't bother investigating it further.
Right now, the OS is newly installed. This is what's happening. Upon running a package install via
apt,timeshift(viatimeshift-autosnap-apt) thinks that another instance is running:If I disable
timeshift-autosnap-apt, packages can be installed, but no other snapshot can be created.Another instance of
timeshiftwas indeed running:This tends to happen when running
timeshiftwith--scripted, but not always. It always happens via script, i.e., crontab,timeshiftscripted or check, etc.mountcannot be killed, even with-9. It's probably an IO error?This is what I see in the logs for subsequent
timeshiftruns:Corresponding
timeshiftlog:Then, it appears to hang there, at
get_mounted_filesystems_using_mtab(), while the hour before, it went further:grep 5958 /var/log/timeshift/*shows:All new attempts at creating a snapshot fail with
another instance is running.No errors on the NVMe itself.
Nothing related to
nvme0n1in syslog, or anything interesting in syslog around that time either:I'm not sure this is related:
Then:
The "weird" thing is there's no trace of these processes in the
timeshiftlogs:Killing the running
timeshiftprocess and executing atimeshiftcommands aftwards hangs:The only way to take further snapshots is to issue a
reboot, followed by a hard shutdown /restart (stay on the power button).