Timeshift snapshots disappeared

All timeshift backups prior to this date just disappeared. Computer was freezing up unexplicably for minutes at a time, then resuming, then freezing again. Only mouse pointer working. When I decided to restore snapshot, noticed the disappearance. Wonder if issues are related.

Already ran “timeshift --check” . It detected a few incomplete snaptshots and removed then, but did not restore my snapshot list.

Any idea what could have happened?

For a backup program, that is a pretty major issue. Thanks !

By the way, is there a way to know if what disappeared has been only the snapshot list or if the snapshots themselves were wiped away?

You can mount the volume where the snapshots are stored at a directory ─ say /mnt ─ and then check whether the snapshots are still there. I suspect that they will be, because it takes a long time for timeshift to delete a single snapshot, let alone all of them.

So it’s probably just the list, but better check, just to be sure. :wink:

Thanks a lot. Pretty newbie here. If that would not be asking too much, I would be glad to know what command line mounts the snapshots. :grimacing:

1 Like

Well, it depends on what partition you are storing your snapshots on. In my case, I keep them on /dev/sdb2, and so on my system, the command would be…

sudo mount /dev/sdb2 /mnt

Then as you look through /mnt, you will find a directory named /mnt/timeshift, and in it, a directory /mnt/timeshift/snapshots. Here’s the output from my own system, picking my most recent backup.

[nx-74205:/dev/pts/3][/root] # ll /mnt/timeshift/snapshots/
total 0
drwx------ 1 root root  94 Aug  2  2019 2019-08-02_05-09-13
drwx------ 1 root root  94 Aug 31  2019 2019-08-31_22-29-05
drwx------ 1 root root  94 May 19 20:07 2020-05-19_20-07-16
drwx------ 1 root root 112 Aug  1 04:29 2020-08-01_04-26-49
drwx------ 1 root root 112 Aug  8 03:43 2020-08-08_03-40-24
drwx------ 1 root root 112 Aug 13 04:38 2020-08-13_04-35-08
drwx------ 1 root root 112 Aug 17 07:31 2020-08-17_07-27-43
drwx------ 1 root root 112 Aug 21 09:27 2020-08-21_09-22-02
drwx------ 1 root root 112 Aug 22 18:25 2020-08-22_18-23-10
drwx------ 1 root root 112 Aug 28 19:17 2020-08-28_19-14-15

[nx-74205:/dev/pts/3][/root] # ll /mnt/timeshift/snapshots/2020-08-28_19-14-15/
total 55004
-rw------- 1 root root     1069 Aug 28 19:14 exclude.list
-rw------- 1 root root      267 Aug 28 19:17 info.json
drwxr-xr-x 1 root root      108 Jun  1 01:34 localhost
-rw------- 1 root root 53407029 Aug 28 19:17 rsync-log
-rw------- 1 root root  2905962 Aug 28 19:17 rsync-log-changes

[nx-74205:/dev/pts/3][/root] # 

The directory localhost in the above terminal output is the actual directory where the files and directories (“folders”) of the snapshot are located.

1 Like

Hello Aragorn,

Mine problem is info.jason file is gone missing whenever I create manual snapshots (haven’t done automatic backups till now).

Also running command “sudo timeshift --list” always shows path at /run/timeshift/backup even when /dev/sdc1 (my backup partition) is mounted in /media/Backups

Could you please help rectify the issue?

This will depend on where the file is stored. By default, cache directories are not backed up. You’ll need to fiddle with the include/exclude settings a bit to get it to backup everything you want, and nothing of what you don’t want.

Yes, timeshift always uses its own mountpoint. This is normal behavior. In other words, you don’t need to have your partition with the backups mounted, because timeshift will mount it by itself automatically. :wink:

Dude, I am such an idiot you know. Initially Timeshift warned me I didn’t have enough space for backup. I ignored it. The second time Timeshift started backing up it didn’t show any warning of partition going out of space. Almost 4 hours of tweaks and googling, I finally looked at rsync logs. Right after /var/lib/libvirt/images , I saw a pipe error. I immediately got an idea that in that images dir I had 40GB space allocated for Virtual Machine(Windows VM), while my partition size for backup was just 35GB. I shouted to myself “What an IDIOT??”

Argh, happens sometimes

And yes thanks for reminding. I missed .cache in backup. Need to dive into Summary and see what I should backup. Also could you tell me which hidden files in home directory is a must backup like .config, .local etc??

timeshift wasn’t designed/intended to back up the home directories, although it can, and in the default configuration if you enable the home directories, hidden files will normally be included already, except for ~/.cache.

1 Like

Okay, I will add ~/.cache in the backup then. One more thing While restoring backup should I first install manjaro from live usb pendrive and then restore or I can directly restore backup without installing manjaro first?

Previously while I was on mint, doing the later procedure as mentioned above everything in home directory was lost (My backup didn’t have home directory backed up). I mean there were no Desktop, Documents, Music, Pictures directory under home directory.

I would recommend installing from the USB drive first and then restoring the backup. That’s how I would do it. :wink:

Did you consider a hard disc defect?

1 Like

Yes absolutely. I though it could be a malfunction. But rebooting into an ext4 or ntfs system would make system work normally. And SMART reports were clean. I have since reformated the partition and installed ext4 and no longer had the problem. I’ve seen other similar complaints about BTRFS.

Either BTRFS has a serious flaw or Manjaro’s implementation of BTRFS is not the best (or not sufficiently fool proof). Considering some large enterprises (like facebook) are making extensive use of BTRFS and some major vendors (like Synology) and distributions (Suse, Fedora) are defaulting to btrfs, I think it might be something with Manjaro’s implementation.

Manjaro does not impart any custom configuration or patches to btrfs. The filesystem driver is exactly how it comes from the upstream kernel developers, and the userland tools to create and manage btrfs come from Oracle, via Arch.

1 Like

A bit disappointing. My info.json got corrupted somehow. The snapshot was still there so timeshift went ahead and deleted the entire snapshot. Dodgy design if you ask me. The info.json doesn’t store mission critical info relating to the actual snapshot.