Unable to upgrade to Dom0 4.3.2: "Failed to connect to system scope bus via local transport"

Updating from 4.3.1 using Qubes Update, getting error. Failed to connect to system scope bus via local transport. Operation not permitted.

Also getting the error @tommy described.

I got the same error. Do you also download your updates over Tor, meaning with sys-whonix? Maybe it’s a whonix problem.

I used sys-whonix as well, after switching to sys-vpn i could update without the error. I don’t know if the error was fixed in the meantime or that it has to do with sys-whonix.

Yep, I was using sys-whonix. Switching worked for me.

Wasn’t fixed, because I still got the same error and I don’t want to switch. I want my updates to go through the Tor network, meaning sys-whonix. So someone has to fix this lol

I get it. But unless you are super targeted, updating plain is not insecure. Using sys-whonix will only help to hide that you use Qubes for your ISP or other observers. But in a public place it should be no problem to update plain (sys-firewall) as proxy. But indeed anyway this issue should be fixed.

I had noticed this in the past, but it was immediately after trying to force-terminate an update by killing various qubes update related processes in dom0 and then rebooting the entire system. I deleted all my Whonix VMs, then reinstalled the templates, created new VMs, and the problem went away, so I chalked it up to user error and moved on.

In order to hopefully fix this problem quickly, because not being able to update dom0 is a big security issue, I want to link this thread on the Whonix forum about the same issue, where more details can be found, and among other things the complete error message, because what have been posted here is incomplete and this is the complete error message:

Failed to connect to system scope bus via local transport: Operation not permitted (consider using --machine=@.host --user to connect to bus of other user)

purging complete whonix and stuff to solve this bug is unfortunate not a solution

As a workaround, you could create a qube that is not based on a Whonix template but still uses sys-whonix for its network, and then set this new qube as the dom0 update proxy. For example:

user@dom0:~$ qvm-prefs default-dvm template
fedora-44-xfce
user@dom0:~$ qvm-create --label red --class DispVM --template default-dvm --prop netvm=sys-whonix update
user@dom0:~$ qubes-prefs updatevm update

That’s kinda useful anyway, cause sys-whonix can be inexplicably slow as a directly configured dom0 update proxy (or maybe just for qvm-template[-gui], which by default also uses the same qube that is configured as the dom0 update proxy).

could one with the error and one without provide output of
qubes-prefs updatevm

Checking for available updates for dom0...
Using sys-whonix as UpdateVM for Dom0
Checking for dom0 updates...
Available updates:
Fedora 41 - x86_64                              2.7 kB/s | 4.6 kB     00:01
Fedora 41 - x86_64 - Updates                    2.8 kB/s | 4.5 kB     00:01
Qubes Host Repository (updates)                 1.9 kB/s | 3.2 kB     00:01

qubes-release.noarch                   4.3-6.fc41             qubes-dom0-current
qubes-release-notes.noarch             4.3-6.fc41             qubes-dom0-current
There are dom0 updates available
Using sys-whonix as UpdateVM for Dom0
Downloading updates. This may take a while...
Failed to connect to system scope bus via local transport: Operation not permitted (consider using --machine=@.host --user to connect to bus of other user)

FYI, unrelated to this specific error but related to updates over Tor, for anyone using the onion repo (http://yum.qubesosfasa4zl44o4tws22di6kepyzfeqv3tg4e3ztknltfxqrymdad.onion/) for dom0 updates, it’s currently offline:

Checking for dom0 updates...
Errors during downloading metadata for repository 'qubes-dom0-current':
  - Curl error (28): Timeout was reached for http://yum.qubesosfasa4zl44o4tws22di6kepyzfeqv3tg4e3ztknltfxqrymdad.onion/r4.3/current/host/fc41/repodata/repomd.xml [Connection timed out after 30001 milliseconds]
  - Curl error (97): proxy handshake error for http://yum.qubesosfasa4zl44o4tws22di6kepyzfeqv3tg4e3ztknltfxqrymdad.onion/r4.3/current/host/fc41/repodata/repomd.xml [cannot complete SOCKS5 connection to yum.qubesosfasa4zl44o4tws22di6kepyzfeqv3tg4e3ztknltfxqrymdad.onion. (4)]
Error: Failed to download metadata for repo 'qubes-dom0-current': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried
Failed to check for dom0 updates

I think @Euwiiwueir is on to something, because it worked flawlessly for me some 12h ago:

user@dom0:~$ sudo qubes-dom0-update
Using sys-whonix as UpdateVM for Dom0
Downloading updates. This may take a while...
[...]
Dependencies resolved.
================================================================================
 Package                 Arch       Version        Repository              Size
================================================================================
Upgrading:
 qubes-release           noarch     4.3-6.fc41     qubes-dom0-current      68 k
 qubes-release-notes     noarch     4.3-6.fc41     qubes-dom0-current      11 k

Transaction Summary
================================================================================
Upgrade  2 Packages

Total download size: 79 k
DNF will only download packages for the transaction.
Downloading Packages:
(1/2): qubes-release-notes-4.3-6.fc41.noarch.rp  16 kB/s |  11 kB     00:00    
(2/2): qubes-release-4.3-6.fc41.noarch.rpm       94 kB/s |  68 kB     00:00    
--------------------------------------------------------------------------------
Total                                            65 kB/s |  79 kB     00:01     
Complete!
The downloaded packages were saved in cache until the next successful transaction.
You can remove cached packages by executing 'dnf clean packages'.
Updating and loading repositories:
 Qubes OS Repository for Dom0                                                                                100% |   0.0   B/s |   1.5 KiB |  00m00s
Repositories loaded.
qubes-release.noarch       4.3-6.fc41 qubes-dom0-cached
qubes-release-notes.noarch 4.3-6.fc41 qubes-dom0-cached
Updating and loading repositories:
 Qubes OS Repository for Dom0                                                                                100% |   0.0   B/s |   1.5 KiB |  00m00s
Repositories loaded.
Package                                          Arch        Version                                          Repository                         Size
Upgrading:
 qubes-release                                   noarch      4.3-6.fc41                                       qubes-dom0-cached             223.3 KiB
   replacing qubes-release                       noarch      4.3-4.fc41                                       qubes-dom0-cached             222.9 KiB
 qubes-release-notes                             noarch      4.3-6.fc41                                       qubes-dom0-cached             794.0   B
   replacing qubes-release-notes                 noarch      4.3-4.fc41                                       qubes-dom0-cached             794.0   B

Transaction Summary:
 Upgrading:          2 packages
 Replacing:          2 packages

Total size of inbound packages is 72 KiB. Need to download 72 KiB.
After this operation, 413 B extra will be used (install 224 KiB, remove 224 KiB).
Is this ok [y/N]: y
[1/2] qubes-release-0:4.3-6.fc41.noarch                                                                      100% |   3.5 MiB/s |  64.4 KiB |  00m00s
[2/2] qubes-release-notes-0:4.3-6.fc41.noarch                                                                100% | 272.9 KiB/s |   7.4 KiB |  00m00s
-----------------------------------------------------------------------------------------------------------------------------------------------------
[2/2] Total                                                                                                  100% |   2.1 MiB/s |  71.8 KiB |  00m00s
Running transaction
[1/6] Verify package files                                                                                   100% | 400.0   B/s |   2.0   B |  00m00s
[2/6] Prepare transaction                                                                                    100% |  16.0   B/s |   4.0   B |  00m00s
[3/6] Upgrading qubes-release-notes-0:4.3-6.fc41.noarch                                                      100% |   5.1 KiB/s | 100.0   B |  00m00s
[4/6] Upgrading qubes-release-0:4.3-6.fc41.noarch                                                            100% |   5.7 KiB/s | 100.0   B |  00m00s
[5/6] Removing qubes-release-notes-0:4.3-4.fc41.noarch                                                       100% | 200.0   B/s |   2.0   B |  00m00s
[6/6] Removing qubes-release-0:4.3-4.fc41.noarch                                                             100% |  34.0   B/s |  24.0   B |  00m01s
>>> Running %posttrans scriptlet: qubes-release-0:4.3-6.fc41.noarch                                                                                  
>>> Finished %posttrans scriptlet: qubes-release-0:4.3-6.fc41.noarch                                                                                 
>>> Scriptlet output:                                                                                                                                
>>> Running timer as unit: run-r743939128a014fbca02852ef5abd8136.timer                                                                               
>>> Will run service as unit: run-r743939128a014fbca02852ef5abd8136.service                                                                          
>>>                                                                                                                                                  
Complete!

So… happy days! :man_shrugging:

I don’t think it’s just a repo problem, because:

  1. You are literally the only one that said that it worked for him with sys-whonix and you never mentioned that you ever had this problem to begin with. So something is fishy here. Maybe you have a highly configured QubesOS that somehow does not have this problem. Or maybe relatively fresh install of QubesOS and maybe only older installs (but that are still updated every day) have this problem.
  2. If it’s just a repo problem, then why does it take so long for someone official from either QubesOS or Whonix to confirm this/relay this information to us?
  3. When I put the error message into AI and read what it could be, it didn’t seem like it’s a repo problem but more like a privilege issue, policy issue or something like that.

Hopefully it’s ā€œonlyā€ a repo issue, but does not seem like it.

There is a new update available for whonix gateway and whonix workstation. After I installed it, everything worked fine again and I could update dom0.
Can you guys confirm that?

Also who is to blame here? As an uncivilized POS I need to put the blame on someone so that I feel better :joy:
But since a Whonix update fixed it and the problem with updating was only through sys-whonix, seems like I need to be uncivilized towards the Whonix project …just kidding of course. But really, what was the problem…what exactly was fixed?

Likewise. Apparently updating helper-scripts 3:55.3-1 -> 3:55.4-1 in the Whonix gateway did the trick.

For anybody still curious, /usr/lib/qubes/qubes-download-dom0-updates.sh returned the Failed to connect to system scope bus via local transport: Operation not permitted error on line 172. Dbus connection failed seemingly due to user id mismatch caused by fakeroot.

More specifically, as @adrelanos have pointed out in the whonix forum thread, leaprun attempt to ensure that tor is running caused the error.

And the ā€œfixā€ is to ignore failed systemctl access: fix: qubes-dom0-update over sys-whonix Ā· Kicksecure/helper-scripts@1265aa8 Ā· GitHub

Both fakeroot and leaprun check have been there for a while at this point, thus I can’t simply write them off as the cause of the issue. If anyone knows how this could work in the first place, or what have caused it to stop working, add to the thread. It is an open mystery for me.

Seems to be corrected for me also. However, I’ve seen ā€˜RPM transaction failed’ during cleanup for the past few months. Not sure what to make of it but I never took it seriously.

[6/7] Removing qubes-release-0:4.3-4.fc 100% | 33.0 B/s | 24.0 B | 00m01s

>>> Running %posttrans scriptlet: qubes-release-0:4.3-6.fc41.noarch

>>> Finished %posttrans scriptlet: qubes-release-0:4.3-6.fc41.noarch

>>> Scriptlet output:

>>> Running timer as unit: run-r1b622dfb4cfc476798c50557591929b2.timer

>>> Will run service as unit: run-r1b622dfb4cfc476798c50557591929b2.service

Transaction failed: Rpm transaction failed.