Error: failed to prepare transaction (conflicting dependencies)

Support System Updates
1

I just tried to run “sudo pacman -Syu” on my '12 Mac Pro edition of Manjaro, which I believe I just updated it last week, today 400 some packages to upgrade “failed” a number of times due to error, “nothing upgraded” and I tried it a few times.

There were some questions about changing packages, which usually in Manjaro the answer is “yes” but I missed the insight on changing from “nvidia570” to “nvidia580” which I think my nvidia card doesn’t do 580 . . . and that was the third “fail” . . . so I ran it again but that time I said “no” for changing and it still “failed”??? Not sure if my machine has reached its lifespan with Manjaro or are there any moves to make to try to fix it? I am running nouveau, but still seems like “nvidia” tries to assert control over the system?

:: Starting full system upgrade...
:: Replace linux70 with core/linux-meta? [Y/n] y
:: Replace linux70-broadcom-wl with extra/linux-broadcom-wl-meta? [Y/n] y
:: Replace mhwd-nvidia-570xx with extra/mhwd-nvidia-580xx? [Y/n] n
:: Replace mhwd-nvidia-575xx with extra/mhwd-nvidia-580xx? [Y/n] n
resolving dependencies...
looking for conflicting packages...
:: mhwd-nvidia-580xx-580.173.02-3 and mhwd-nvidia-570xx-570.211.01-1 are in conflict. Remove mhwd-nvidia-570xx? [y/N] n
error: unresolvable package conflicts detected
error: failed to prepare transaction (conflicting dependencies)
:: mhwd-nvidia-580xx-580.173.02-3 and mhwd-nvidia-570xx-570.211.01-1 are in conflict

$ lsmod
Module                  Size  Used by
rfcomm                110592  4
qrtr                   57344  2
squashfs               98304  0
loop                   45056  0
bnep                   36864  2
btusb                  86016  0
btmtk                  32768  1 btusb
btrtl                  32768  1 btusb
btbcm                  24576  1 btusb
btintel                73728  1 btusb
bluetooth            1196032  28 btrtl,btmtk,btintel,btbcm,bnep,btusb,rfcomm
mousedev               28672  0
apple_mfi_fastcharge    16384  0
vfat                   28672  1
fat                   114688  1 vfat
intel_powerclamp       24576  0
snd_hda_codec_alc882    20480  1
snd_hda_codec_realtek_lib    65536  1 snd_hda_codec_alc882
snd_hda_codec_nvhdmi    16384  1
snd_hda_codec_hdmi     61440  1 snd_hda_codec_nvhdmi
snd_hda_codec_generic   114688  2 snd_hda_codec_realtek_lib,snd_hda_codec_alc882
coretemp               24576  0
snd_hda_intel          73728  5
kvm_intel             536576  0
snd_hda_codec         221184  6 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec_nvhdmi,snd_hda_codec_realtek_lib,snd_hda_codec_alc882
wl                   6524928  0
snd_hda_core          147456  5 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec,snd_hda_codec_realtek_lib
kvm                  1474560  1 kvm_intel
e1000e                397312  0
snd_intel_dspcfg       49152  1 snd_hda_intel
applesmc               32768  0
snd_intel_sdw_acpi     16384  1 snd_intel_dspcfg
i2c_i801               40960  0
snd_hwdep              24576  1 snd_hda_codec
irqbypass              16384  1 kvm
iTCO_wdt               16384  0
snd_pcm               221184  4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec,snd_hda_core
cfg80211             1507328  1 wl
intel_cstate           20480  0
intel_uncore          282624  0
acpi_cpufreq           32768  0
intel_pmc_bxt          16384  1 iTCO_wdt
i2c_smbus              20480  1 i2c_i801
i2c_mux                20480  1 i2c_i801
pcspkr                 12288  0
ptp                    53248  1 e1000e
snd_timer              57344  1 snd_pcm
snd                   159744  18 snd_hda_codec_generic,snd_hda_codec_hdmi,snd_hwdep,snd_hda_intel,snd_hda_codec,snd_timer,snd_hda_codec_realtek_lib,snd_pcm
pps_core               32768  1 ptp
i5500_temp             12288  0
soundcore              16384  1 snd
ioatdma                81920  0
i7core_edac            45056  0
rfkill                 45056  6 bluetooth,cfg80211
mac_hid                12288  0
dca                    20480  1 ioatdma
crypto_user            16384  0
uinput                 28672  0
dm_mod                233472  0
nfnetlink              20480  1
nouveau              3784704  10
drm_ttm_helper         20480  1 nouveau
ttm                   126976  2 drm_ttm_helper,nouveau
video                  81920  1 nouveau
gpu_sched              73728  1 nouveau
i2c_algo_bit           24576  1 nouveau
drm_gpuvm              57344  1 nouveau
firewire_ohci          69632  0
drm_exec               12288  2 drm_gpuvm,nouveau
sr_mod                 32768  0
mxm_wmi                12288  1 nouveau
hid_apple              32768  0
firewire_core         253952  1 firewire_ohci
wmi                    36864  3 video,mxm_wmi,nouveau
cdrom                  86016  1 sr_mod
crc_itu_t              12288  1 firewire_core
drm_display_helper    290816  1 nouveau
lpc_ich                28672  0
cec                    98304  1 drm_display_helper
2

This is a message you’ll get when you’re running an EOL kernel. Kernel 7.0 is EOL, and linux-meta will then pull in the next still supported kernel with a version closest to what you were running before.

Presumably this is also why you are running into trouble with the Nvidia drivers.

3

OK, thanks for the reply, that “replace linux70” question might have been one of the questions that I said “yes” to . . . running on “automatic” . . . as my newer homebuilt machine doesn’t have nvidia card, so jumping back and forth might have gotten me into trouble.

Any way to resolve the problem or “fait accompli” the / is now “toasted”???

4

I would recommend installing either kernel 6.18 LTS or kernel 7.1, and once booted into that, remove and reinstall mhwd. It’s the quickest way to solve the problem. :thinking:

7

Interestingly, in my new '25 or '24 ryzen machine my Manjaro install is running the 6.12 kernel . . . I keep forgetting that Manjaro is a “manual upgrade” on the kernel?? It is one of 8 installs, so it doesn’t get rapt attention.

Anyway, flipping back over to the Mac Pro in a minute to check where it is kernel wise.

Alrighty, the saga continues . . . got the kernel boosted to 7.1, tried to remove 7.0 in the GUI app and it says “Done,” but nothing happened. Removed an earlier 6.6 kernel and it shows that it doing stuff. Anyhow, booted in 7.1, tried to remove “mhwd” but it declines due to conflict with “mhwd-nvidia” tried a few different moves including removing “mhwd-nvidia580” . . . but, none of those moves worked out in the GUI handler.

Went back over to the console and ran the sudo pacman -Syu again and now back to this question:

sudo pacman -Syu
[sudo] password for noone00: 
:: Synchronizing package databases...
 core is up to date
 extra is up to date
 multilib is up to date
error: failed retrieving file 'core.db' from ridgewireless.mm.fcix.net : Connection timed out after 10002 milliseconds
error: failed retrieving file 'extra.db' from ridgewireless.mm.fcix.net : Connection timed out after 10002 milliseconds
error: failed retrieving file 'multilib.db' from ridgewireless.mm.fcix.net : Connection timed out after 10002 milliseconds
warning: too many errors from ridgewireless.mm.fcix.net, skipping for the remainder of this transaction
:: Starting full system upgrade...
:: Replace linux70 with core/linux-meta? [Y/n]

Should I say “no” to this one or “yes”???

[edit] Ah, I got this question now that 70 is EOL we do want to move to “core/linux-meta”??

[edit2]

:: Replace linux70 with core/linux-meta? [Y/n] y
:: Replace linux70-broadcom-wl with extra/linux-broadcom-wl-meta? [Y/n] y
:: Replace mhwd-nvidia-570xx with extra/mhwd-nvidia-580xx? [Y/n] n
:: Replace mhwd-nvidia-575xx with extra/mhwd-nvidia-580xx? [Y/n] n
resolving dependencies...
looking for conflicting packages...
:: mhwd-nvidia-580xx-580.173.02-3 and mhwd-nvidia-570xx-570.211.01-1 are in conflict. Remove mhwd-nvidia-570xx? [y/N] n
error: unresolvable package conflicts detected
error: failed to prepare transaction (conflicting dependencies)
:: mhwd-nvidia-580xx-580.173.02-3 and mhwd-nvidia-570xx-570.211.01-1 are in conflict

Can I remove “mhwd-nvidia-580” via some console command, Or should I run back through and see if the upgrade will install 580???

OK, back to where we were with the first failure, answered yes to all of the questions, and ran it through, as before, back to something about “lightdm-gtk-greeter”??

(411/411) checking for file conflicts                   [##############################] 100%
error: failed to commit transaction (conflicting files)
lightdm-gtk-greeter-settings: /usr/lib/python3.14/site-packages/lightdm_gtk_greeter_settings/__pycache__/installation_config.cpython-314.pyc exists in filesystem
Errors occurred, no packages were upgraded.

Mod Edit: Fixed formatting / merged consecutive posts

8

I’m not sure if that should be part of the package, which is presumably why the error occurs (as it wasn’t previously owned by the package).

Anyway this should get you past the error.

sudo pacman -Syu --overwrite="/usr/lib/python3.14/site-packages/lightdm_gtk_greeter_settings/__pycache__/*"

If the files in __pycache__ get removed from the package later, then an update should remove them and they should be replaced when the software is run.

9 
sudo pacman -Rns mhwd mhwd-db && sudo pacman -S mhwd

Refresh your mirrors first before doing anything else. :backhand_index_pointing_down:

sudo pacman-mirrors -f
10

Thanks to you both for the suggested commands, for the most part I haven’t had issues with running Manjaro, so I don’t have any console chops for digging myself out of problems with it.

Work day today, have to get back to this in a few days . . . I’ll post back with any relevant data. Or mark as solved.

13

OK, had a minute to get back to this, so I did firstly refresh the mirrors . . . .

And, then I ran this “overwrite” command and that led to the pacman -Syu and the four questions, which I answered all of them “yes” and that led to a 411 package upgrade and it showed in there “removing mhwd xxx” for a couple of packages, and it ran through OK, even though there some “errors” showing in that process for one thing or another, it returned to the cursor w/o error.

Point being I didn’t have to run aragorn’s suggested “-Rns” command, because that got picked up in the overwrite, but I don’t know enough to know if it was the refresh mirror that was the “solution” or if it was the “overwrite” command. But for now (before rebooting) it seems the upgrade went through. Issue remains is this aging machine capable of running nvidia580?? because I believe in TW it does not?? Posting this before I reboot and update the grub bootloader to then get back here to Manjaro.

[edit] Back in the biznez . . . reboot and log in is workinng, where there has been some issue is in reviving from suspend as the software moves ahead of the nvidia hardware, but no real worries as this machine isn’t the daily driver . . . . Have to check that later. But for now thanks for that assistance . . . for the most part Manjaro is trouble free, no extra skills needed to keep it going, etc. [edit]

14

Any errors should be dealt with. What were they?

paclog --after=2026-07-21 | paclog --warnings

You can adjust the date as needed. To view the entire log just use paclog.

Refreshing mirrors won’t fix an exists in filesystem error, unless you get a different version of the package that doesn’t include those files.

This is how you can double check:

pacman -Ql lightdm-gtk-greeter-settings | grep pycache

If it lists any files then they’re included in the package.

15

Thanks again for sharing those commands. There was no “W” “error” line, it was just something in the -Syu process that showed “error” . . . . possibly in some checking of one of the kernels.

Back in the MacPro today, and in Manjaro, ran another -Syu and it showed 145 packages to upgrade, must have been a new kernel because it went through os-prober . . . there was the word “error” showing in the processing, but no “W” error and they all upgraded fine. So, back in smooth waters?

16

What’s a ‘“W” error’?

I assume you mean it didn’t show up in the filtered logs.

We can’t access your computer, we have to rely on what you tell us. Try reading things and if in doubt then paste it here so we can see it.

You say os-prober was run so it could be this:

Root filesystem isn't btrfs
If you think an error has occurred, please file a bug report at "https://github.com/Antynea/grub-btrfs"

If so then it’s not an error, just information (assuming / isn’t btrfs).

Probably, but my :crystal_ball: is broken. :smiley:

Sounds like things are ok.

17

Yes, seems like things are again “OK” in that install of Manjaro, so I didn’t need to look at the logs for it.

A “W” error is what shows up in my Debian/Ubuntu and I think even my SUSE installs if there is some hard error that interferes with completing the dist-upgrade, or as with my recent experience with Manjaro where the -Syu process was run, but errored out, could not be completed . . . but each time, rather than saying up front, “This ain’t gonna work, so I’m not doing it.” It ran through package retrieval and “install” until the end . . . citing the “lightdm + paycache” problem.

In Debian type installs running an “apt dist-upgrade” it will show:

W: we couldn't find /.
W: this is a problem.
xxxx@localhost: <blinking cursor>

It seemed like way back that might have shown as “E” for “error” one guesses. Sometimes the system will offer a potential fix, “you might try doing this or that.” But, sometimes it just shows the error, figuring that that would be enough to clue the end user into the fix, etc.

In this recent running, it probably might have been something like your suggestion of “/ isn’t btrfs” because I don’t run it, but if it was something like that I wouldn’t bother mentioning it . . . it was some other superfluous error, but not enough to flag it for a “W” type error line.

18

Debian/Ubuntu use apt/apt-get (there is W: and E:). W stands for warning and E for error, IIRC,

19

I haven’t used apt-get in a long long time and I don’t remember any errors like that, so I must have been lucky (or forgetful).

We just call them errors or warnings.

In future please show us so we can see what you’re talking about, otherwise there’s no point mentioning it in the first place.

20

Alrighty, “nothing to see here” . . . it was mentioned because at that time the pacman process wasn’t completing smoothly . . . . But that is now in the rear view mirror . . . .