I did install a clean qubes 3.1 and I created all vms by myself. After a while I noticed some weird behavior with the autostart setting in qvm-prefs.
My sys-net and sys-firewall were auto starting even with autostart set to False, and so I though the system must have written somewhere to start these machines since the name I've choosen is the same the system usually name these machines.
Decided to change the names and even create new net and firewall vms to see if both network and firewall kept auto starting.
After a while I can't really say why they start or not. Definitely qvm-prefs autostart is not helping.
Right now I've to machines with name network and firewall which keep autostarting even with qvm-prefs autostart set to False.
Something else I found out is the state of the vms in qubes manager when vms are set to autostart. They do not fully start. They remain in that yellow state (starting). But only until you launch an application from the vm it won't fully start (with green state). For example, in network vm only until I launch a terminal It won't start and I won't get NetworkManager Applet.
Something weird that happened a few times is when launching an application, the vm "start" and the application launch but the qubes manager shows again a yellow state in that vm and next to it, a recycle symbol (green arrows) like when you update a template and a vm need to be restarted.
Another issue I've found is after setting the debian 8 template. I tried to create a vm but I was receiving the following error:
Dom0: Error Creating AppVm
Error: 'NoneType' object has no attribute 'qid'
Also, while trying to change the netvm of the Template was encountering few problems. Qubes manager showed a netvm but when tried to change it from qubes manager I could not, or changes were not applied or different error messages were showed. (the only thing I remember is mentioning qid in various of these errors.
Since I couldn't do much decided to continue setting fedora 23 template and after 1 or 2 reboots the debian8 template issues disappeared and I could create and start vms from debian.
Last thing I wanted to mention is my inability to use the "convert to trusted PDF". I did mention this in another post. The application starts, open a disposable VM and send to Untrusted PDFs the original one, but there is no secure PDF as a result.
I know most of this bugs must be minor bugs between manager and vms but I do not know how to help it. I wish I could.
cheeers!
-> Backing up files: 94%...Removing file dom0-home/User.004
-> Backing up files: 94%...Sending file dom0-home/User.004.hmac
-> Backing up files: 94%...Removing file dom0-home/User.004.hmac
-> Backing up files: 94%...tar: User/.local/share/systemd/user: File removed before we read it
-> Backing up files: 94%...Wait_backup_feedback returned: addproc
ERROR: Failed to perform backup: error in addproc
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Tue, Dec 15, 2015 at 06:15:15AM -0800, Cube wrote:
> I'm consistently getting the same error and can't get around it, here's the
> error
>
> -> Backing up files: 94%...Removing file dom0-home/User.004
> > -> Backing up files: 94%...Sending file dom0-home/User.004.hmac
> > -> Backing up files: 94%...Removing file dom0-home/User.004.hmac
> > -> Backing up files: 94%...tar: User/.local/share/systemd/user: File
> > removed before we read it
> > -> Backing up files: 94%...Wait_backup_feedback returned: addproc
> > ERROR: Failed to perform backup: error in addproc
> >
>
> Before you think the "File removed before we read it" is the issue, I've
> tried in multiple ways to delete that file and it still dies at the same
> point, so that doesn't seem to be the problem. Any ideas? Thanks much ...
But if you've tried to remove that file, why it is still there?
Did you get any other error then?
> then rebooted and tried again (maybe it was some odd problem), this
> is the result of that so shows that error again.
I see that .local/share/systemd/user is indeed broken symlink. You can
try to create the directory pointed by it (.config/systemd/user). As the
symlink is apparently recreated after reboot...
Anyway a single broken symlink shouldn't break the whole backup:
https://github.com/QubesOS/qubes-issues/issues/1515
And the final error is still about "addproc"? If there is no other
message, I think it's time to go for --debug option. Be prepared to
receive A LOT of output.
Oh one thought - "why is it still there"? No idea, but seems to get recreated, at least on reboot.
The user name on this is "User" (User@dom0), and this file is named "user", perhaps some issue with using the name "User"? I remember you mentioning something about that.
Now tht you mention it I notice this too. I k ow I get the message when I start the ws.client vm. I do recall reading something in whonix sites docs that the newest version is now using silent mode for the connection and bootstrapping of the gw. Not sure if that has anything to do with it. I can not recall what was needed to change it back to.the gui verbose pop up window for it. Maybe see if you can find it on their site if you think it might be the reason for this.
Another thing I noticed was a difference in some of the config of vms created by the command line versus the vmm gui. I can't recall off the top of my head what they were. I k ow it was enough for me to delete them and ceate them instead via the vmm. I will look at this when I get home
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Tue, Dec 15, 2015 at 04:18:55PM -0800, Cube wrote:
>
>
> > And the final error is still about "addproc"? If there is no other
> > message, I think it's time to go for --debug option. Be prepared to
> > receive A LOT of output.
> >
> >
> Yes, this IS the result of --debug :)
Ah, indeed...
> Anything else I can try?
Is that still error in "addproc"? If so, there is some error message there,
maybe much earlier...
- --