"pamac upgrade" just terminates and breaks my system

Support System Updates
1

It’s not the first time when it happens. I run “pamac upgrade” and at some point while updating it just terminates saying like “update was terminated.” Ive noticed it usually happens when i do anything at all while updating for example i ran a “find” command in another terminal or closed my browser. If i run it at fresh boot and dont touch anything - it goes without problems.

The main problem - when it terminates the update like that, it breaks my system (also i cant run the update again cuz of “waiting for another package manager…”). After this termination i reboot my PC and see that none of my kernels are found (except 5.14 which i cant even find installed in my system but it works when i boot it (only TTY tho)). I reinstall the kernel, boot the system and it seems to be working.
But when i run “pacman -Qkk” it shows a lot of errors either “change time does not match” or “size does not match”.
The last times it happened i ran the command to reinstall every single package pacman knows and it seemed allright. Now it seems ill have to do it again to fix this, but my main problem is:

Why does pamac just terminates my updates like that when i do anything? It actually terminates at the “Downloading packages” stage, not the “Installing packages” stage.
Pacman (yay actually) has no such problem, i update and reinstall with it.

1 Like
2

Have a look here:

3

it really looks similar to my problem. i also got this error in pacman after booting, i just removed the directory and did Syyu (and reinstalled onlyoffice with “–overwrite” flag (which told me permissions are different: system has 775 and the package wants 755)) which fixed the issue

error: could not open file /var/lib/pacman/local/onlyoffice-[something]/desc: No such file or directory

4

Yes, I know, the installer default to using btrfs, what can I say?

I am thinking - are you using btrfs as filesystem?

I think something else is causing the failure - for the past couple of weeks I have been testing my upgrades using pamac upgrade --no-aur on the command line - without any issues.

My chosen file system is ext4 so if you are using btrfs I am thinking that might be the culprit.

5

i use ext4 aswell. and i was using “–no-aur” this time and it terminated as always.

6

looks good

That is strange - my updates pass without issues.

Just now I updated one of my servers over SSH, using the exact command, no failure - pure success.

EDIT:
Just to see if I can reproduce your issue, updated using pamac upgrade

  • my primary laptop (ThinkPad X13)
  • my secondary laptop (Tuxedo Infinitibook Pro 14 gen8)
  • my gaming laptop (Lenovo Legion with Nvidia 5070)
  • my workstation (Lenovo ThinkStation P620)
7

to try to reproduce it you should do something while the update. for example run the “find” command in another terminal in graphics or kill some apps like browser

8

I still cannot reproduce the issue.

Manjaro has die hard GUI users which use the Pamac GUI without to do major updates without any issues.

I am thinking - perhaps you are running out of disk space or swap is not configured - in any case - it seems to be something local to your system.

1 Like
9

well, apparently im not the only one with this problem:

ive got a lot of free space and a lot of RAM so i dont have a swap.

1 Like
10

I believe you, that you experience an issue, no doubt, I am not questioning that.

Yes, I have seen a couple of topics - but only a few, there is nothing that signals a general error with pamac or pacman (which provides the libalpm used by pamac).

In the topic you reference - even pacman crashed - which makes me think there is more to it than just pamac crashing.

Technically it could be libalpm that causes the crash, but then it would be a high priority issue on the Arch Linux BBS.

I have tried to reproduce the issue on five (5) different systems - not just today - and had no success in getting the pamac commandline tool to fail.

2 Likes
11

My pamac problem is not so bad as yours but I am sick of every time I do a search it totally locks up my system. I admit I keep forgetting to use cli mode, but this needs to be fixed.

12

The forum has many threads these days about problems with pamac, to such an extent that I’ve included it in the “Known Problems and Solutions” section of the 2026.02.23 Stable Update thread, with the advice to update your system with pacman or with octopi.

But unfortunately, people just seem love doing things that end badly. Or, they can’t read. Or both.

:man_shrugging:

3 Likes
13

The forum has many threads these days about problems with pamac, but unfortunately developers dont seem to care to fix them or at least inform about them not only for people who read all problems on forums of a distro which is supposed to be user-friendly
(also forum has a lot of messages about problems with updating with pacman cuz of dependencies and it’s recommended to use pamac in there)

14

The situation is actually a little more nuanced than that. @guinux, the maintainer of pamac has been alerted about the issue, but the problem is that even though he is still the official maintainer, he’s not as tightly connected with Manjaro anymore as he used to be — possibly because of professional or other obligations.

As for why there was no official information about the problem, that’s another story, and again, not as black and white as you put it.

1. Problem exposure

Even though there has already for a while been an occasional instability in pamac — and I’ll be the first to admit that it contains many more bugs than Arch’s venerated pacman or the very decent octopi — the problem we are seeing now, with the pamac GUI crashing in the middle of an update, with or without that the pamac daemon continues updating the system in the background without any visual feedback, is still fairly new.

I’m no developer, but I suspect that it could be related to the recent changes to polkit, given that there weren’t any other incisive changes in that regard. As such, not everyone on the team was perceiving the problem as particularly indicative of a serious flaw just yet.

In addition to that, not everyone on the team actually uses pamac — least of all the pamac GUI — for updating their systems.

2. Misguided pride, perhaps?

pamac is one of Manjaro’s own creations, and therefore, a halo product. As such, I can imagine that it would be a bit difficult for the developer team to admit that Manjaro’s halo product is seriously flawed, and especially so given the nowadays less integrated role of pamac’s maintainer within the team.

3. A heads-up has effectively been issued

The “Known Problems and Solutions” section of the 2026.02.23 Stable Update thread does now contain a warning about the problem, with the advice to either update one’s system from the command line with pacman, or to use octopi for those wanting a reliable graphical package manager front-end — octopi directly uses pacman at the back-end, in combination with an AUR helper (such as yay) if so desired.

I have added this warning myself, after it became clear that there was something seriously amiss with pamac, and that — at least, from my own perspective — the crashes have become (more) prominent since the 2026.02.23 Stable Update.

This is not true. pacman handles dependency problems just fine. The problem you are referring to is caused by the proprietary Nvidia drivers and their dependencies, not by pacman, and if people were to actually use their brain, then they’d know how to handle that.

If you cannot remove a package or update a package because of dependencies — and again, Nvidia is the only area where this problem occurs — then you must remove the dependencies first. It’s not rocket science.

5 Likes
15

im not referring to NVIDIA problem, i agree with you on this case. im referring to some old troubles since which i was told to use pamac.

16

You were probably told to use mhwd, although I will concede that pamac does have its fanboys among the members, and that they would rather advise you to use that than to use the tried-and-tested, and nearly perfect pacman.

1 Like
17

Not sure where I fit there. But I have had No issues with pamc since installing Manjaro KDE on multiple computers … 4 now, and one of them is currently running via a USB connected SSD.

Not with pamac-manager, not with pamac-cli. My partner uses pamac-manager, because she is unwilling to learn to use the CLI. It has not bitten her once.

Does that make me a ‘fanboi’?

The closest I have come to an issue that might have been blamed by others, on pamac, was when I got consistent failures to to update. At that point, if I was a less experieced Linux user, or someone who doesn’t understand how to break a problem down into smaller steps, I might have have posted a thread about pamac-manager failing during an update.

Instead I checked the psuedo Terminal in pamac-manager, and lo and behold, the issue was actually a REpo server failing to respond correctly, to a request. The error message was something like 'Slow Server response.." and some miniscule data transfer rate.

Changing to a different set of Repo servers fixed the pamac-manager ‘issue’.

2 Likes
18

No, I was not thinking of you when I wrote that. I was rather thinking of people always offering their advice using pamac commands rather than pacman commands. :wink:

1 Like
19

Actually rather less so than @Aragorn would fit the Octopi fanboi label :wink:

The overarching killer feature of pamac is that it isn’t limited to pacman/AUR upgrades…

If it launched it’s processes in a real terminal, then I’d be more of a fanboi… this is one thing I enjoyed about octopi.

2 Likes
20

Yes having a real terminal like Octopi does would be good. Also making the current psuedo terminal access more obvious would also be good, so too would making those error messages available to the user, instead of a message saying the Install/upgrade failed, because that is the only indication I got, that something went wrong.

1 Like