Makes total sense, the problem is when you remove your Ophans and update instant after it, without using all program’s and all possible features around them.
You will probably never know if the program’s don’t work after, because of the update or because you removed orphans.
I’m already running a Timeshift backup on an external hdd before every update and save all my personal files with Borg via Vorta GUI. I haven’t tested the waters though which means I don’t know exactly how to restore the Timeshift backup especially without a running window manager. I thought that’s a bridge I cross when needed.
But maybe I really should read up on this realizing how quickly I can run into problems with pacman.
That’s why I have decided to wait with removing the orphans after the update now. 
Was gdm explicitly installed by default after installing Manjaro Gnome? Or did you accidentally mark it as a dependency?
I tried playing a new Manjaro Gnome ISO on KVM/QEMU and see that gdm is explicitly installed by default:
Install Reason : Explicitly installed
I haven’t made any significant changes to my GNOME system since I installed it back in February '23. I’ve investigated the issue in the forum though and found a post by @deemon back in August '23:
I can tell you, that using Timeshift GUI is really really userfriendly atleast on a EXT4 Filesystem with RSync. It tells you exactly what is going to replaced, all the Partitions that will replaced with a good overview (like boot,swap,root) (even home if you manually added it in the rules, but per default it is skipped, so its not get mirrored/restored) and you just press a few times okay, to accept everything, then comes the filecheck window and after that the last agree and its going automatic in tty for you and you just lean back and be surprised how relieable timeshift really is.
It love this GUI and its total userfriendly. But i heard bad stuff about BTRFS and BTRFS Snapshots… but rsync should be fine (at least that what i heard), so my own experience going only around ext4+rsync.
Here some good points from me, how to update:
Because nothing depends on it except gnome-initial-setup. Arch assumes a user would install GNOME during installation with sudo pacman -S gnome which installs all the gnome group packages explicitly.
However, on Manjaro, gdm is installed explicitly so it never should be an orphan. As @linux-aarhus mentioned, you should mark it as explicitly installed.
No, you will not.
That’s encouraging because I also use the combination ext4 + rsync. According to your description it seems to be an easy task restoring my Timeshift snapshots. 
I’ve read your comment days ago because I religiously 
read all the 400+ comments in the Stable Update 13.05 announcement and continue monitoring the 18.05 one for potential upcoming difficulties.
Mine was as you can see by the pacman -Qi output above. I assume it’s because of the removed ‘meta-packages’ @deemon describes in the other discussion.
Nevertheless, I’ve marked gdm as ‘explicitly installed’ now with
sudo pacman -D --asexplicit gdm
Now I’m asking myself though if there are more system packages in my pacman -Qdt list in the first post which have been orphaned because of this ‘meta-packages’ thing. Packages like xorg-server-xvfb and wayland-protocols read as if they could be system packages. Let’s see if these are still orphaned once I’ve done the upgrade.
Thanks for confirming. I marked ‘gdm’ as explicitily installed (mentioned above) once I’ve read that.
Off topic
Just FYI, there is a worse case: Timeshift (with rsync tool) cannot save you if you or a malicious program accidentally delete or modify data on your system recursively while another backup disk is mounted. Backup and (rsync) snapshots could be gone.
From my understanding the orphan list must be seen with a grain of salt… it shows packages that you may not need, but i wouldn’t count on it.
Even Steam is under orphan’s listed on my system… but i use it daily.
I just go in pamac>install>orphans and show them by size and remove only the biggest packages, after investigation if i need them and ignore the rest.
I mean, who cares if there are small orphan’s on your drive, its more a handicap if they have bigger size because of the next update, ssd wear and ssd freespace because a ssd with a lot freespace is much faster than ssd which is filled to the rim.
When there is another handicap around orphan’s and why they should be removed, i would love to know more about it… my information’s are still limited around this topic.
I’m not sure if i understand your example… with recursively you talking about a repeated process? You mean like a filechange, while you created the timeshift snapshot? I always was thinking about this, thats why i don’t do anything around my system while the Snapshot created… it takes 1-2min anyways. Do you mean this?
Im actually create my Timeshift Snapshot’s on a external usb drive and switched it between my PC and Laptop with 2 seperated timeshift ext4 partitions.
Can you specify, what you mean with it? From my understanding,
timeshift just shows me which partitions are gonna be restored.
A third drive/disk would always be ignored if its not exactly included in the backup progress (which needs manual specific intervention anyways).
Atleast this was my experience specially on my PC with additional HDD’s/SSD’s connected and never showed up in the Timeshift restore progress (never listed for replacement). My mounted NVMe EXT4 (game drive) or my HDD EXT4 (Temp/Download Partition), was always excluded around the Timeshift process.
I never saw that my rsync snapshots are gone, but maybe i miss your exact point how this situation can be trigger. 
No, there are many examples: Run rm -rf ... , chmod -R ... or write your own Python program incorrectly, these could eat up all data, cloud data, rootless backup and writable snapshots if your external hard drive is permanently mounted on your system.
I read that some people have experienced losing their backup due to poor backup solution.
Offtopic
Ahh yeah, now i got you… i just recently was talking about the Layouts on the KDE Store which deleted all home mounted drives:
Lucky i always create my backup drives in addition to external HDD’s, that only connected when i really need them or to refresh my backup again.
My main focus was to evade lightning strikes or overvoltage… so your backup is history. I got recently the wake up call from the Plasma6 KDE Global Theme Story, i wasn’t aware about this dangerous commands that exists around Linux.
Long story short, this has nothing exactly to do with Timeshift per se… just a global problem how easy can your files can be lost, if you not create external backup, that should be seperated from your system’s, if you want to rely on the backup.
Agreed.
Nonsense. 
Offtopic
Im 100% sure about this point… there are maybe a few SSD/NVMe Models which don’t show a big performance decrease impact… but the majority from SSD Models are heavily affected… a empty SSD performance the best, while you see already a impact on 50% fill status (around -30%) and 80% fill status on some models show -30-50% performance degration. (This is my general statement without refreshing my memories).
Just for you the information Source:
TBH i mainly rely on the SSD/NVMe performance tests on computerbase.de (german) because idk which english article do fill status testing (because they are important).
Maybe you can use sitetranslation, Füllstand stays for Fillstatus btw
There are many models shows the maximum Writespeed can hold for a much longer time when the SSD is empty and you can see reduced max MB/s from some models they are filled:
For example my “Performance Writting after fillstate (Corsair MP600 2 TB)” that i own shows from 70seconds max speed while empty, to 35sec max speed while filled at 50% to 15sec max speed while filled at 80%.
The (WD Black SN750 1 TB) shows another picture, it has zero impact on the fillstatus… but as i said, the majority shows impact and i see other SSD tests in the history where SSD decreased to HDD speed after they was filled to 80-90%.
For that reason i always buy SSD/NVMe, that will have atleast 50-60% room, to have Top Performance… Im optimising my PC’s since 486er times.
There’s no word of warning about this in the Wiki about System Maintenance.
I tried to remove a couple of packages over the past few days with sudo pacman -Rs … and almost always would have removed some important optional dependencies for installed packages. Earlier I would have also removed the GNOME Display manager but luckily @Zesko warned me before doing it.
I think a little note/warning in the Wiki might be a good idea so that users don’t end up with borked installations just because they followed Manjaro Wiki guidelines.
Not possible with the commands provided at the wiki.
And it does suggest running -Qdt first and inspecting the list before removing.
(which is itself not really necessary as the removal action will still stop and prompt)
I think pacman -Rs $(pacman -Qdtq) could lead to the removal of optional dependencies because of the --recursive flag. I recall a warning being given, so, just as you point out, one should always pay attention to the output when running these commands.
Where in the pacman part does it suggest to ‘inspect’ the list’? Nevertheless, I don’t want to rule out I’m misunderstanding the process. Let me ask then - if I
sudo pacman -Rs $(pacman -Qdtq)
and pacman warns about optional dependecies like:
checking dependencies...
:: hplip optionally requires python-reportlab: for pdf output in hp-scan
:: timeshift optionally requires xorg-xhost: For authorization on Wayland
…
would pacman remove these packages (like python-reportlab) if I continue or is it telling me that it won’t uninstall those because those are still optionally required?
On a sidenote - when I tried to remove some of the packages which pacman -Qdt returns me (see first post), pacman regularly warns me about them being ‘optionally required’ by some other package. What command is pacman executing here to get this information?
I know that pacman -Qi python-reportlab returns this information in one of its lines labeled ‘Optional For’ but what is pacman executing to get just this little piece of information for output in a pacman Rs command?
This topic was automatically closed 36 hours after the last reply. New replies are no longer allowed.