First I lowered Private storage max size of some qubes
Then when I tried to start other qvms I noticed them failing and shutting down with this error.
So I decided to just restore my system to before resizing with on one of the two backups I just made in the past week. But then I get the following error:
sys-firewall failing to start:
Error 60 seconds into starting sys-firewall:
Cannot connect to qrexec agent for 60 second,
see /var/log/xen/console/quest-sys-firewall.log for details
In that log thers a lot of stuff… Here are random lines I am hoping are relevant:
GPT:Primary header thinks Alt. header is not at the end of disk.
GPT: 4### != 8###
GPT: Alternate GPT header not at the end of disk.
GPT: User GNU Parted to correct GPT errors.
...
mount: can't read '/proc/mounts': No such file or directory
...
systemd[1]: Listening on systemd-oomd.socket -Userspace Out-Of-Memory (00M) Killer Socket.
...
Finished. [0;1;39mmodprobe@loop.service.[0m -Load Kenel Module loop.
Job qubes-rootfs-resize.service/start running (3s /no limit)
...
Job qubes-rootfs-resize.service/start running (57s /no limit)
If I “Open Console in qube”:
Job qubes-rootf-resize.service/start running (35s / no limit)
journal ctl -f shows this in red:
device not found no device found at address '(<null>)' matching MAC address '....' and alias '(<null>)'
Backup Restore Errors:
Restore error!
Corrupt backup header (hmac varification failed)
Is the password correct?
I’m pretty sure there’s no password issue.
So I tried the other and same issue.
After making both of those I “Test restore to verify backup” and got no errors.
When testing the backup, do it on the last media it will be backed up from.
I messed up by testing it in a qvm vault then moving it onto my external media.
Going forward I have avoided issues by backing up directly to my external media.
Resizing:
Don’t shrink! To avoid this, I created a clone of the qube that has more storage and one that has less.
Suggestion to Qubes Devs: Add warning message in GUI about shrinking. I think many users trust the GUI wont break stuff this severely, and a simple pop up warning & confirmation might save a lot of pain.
Final “FIX”:
Backed up customized templates (which was the bulk of my config work) directly to the external media. Then I freshly installed QubesOS then restored those templates and manually created all my custom qubes.
I think It would work to also backup all the qubes except for the one that was shrunk that created the resize error, but I did not try that.
Using salt, a downsize fails with a helpful message.
At the CLI, shrinking will indeed fail with a warning message, and will
only work with the --force option.
The GUI simply will not process a shrinkage. That is, if you have
private storage set at 20GB and attempt to shrink, the smaller figure
will be ignored. Agreed, there’s no warning that the effort to shrink has
been ignored, or why, but it wont be processed. How exactly did you
manage to shrink the volume?
I never presume to speak for the Qubes team.
When I comment in the Forum I speak for myself.
Hi @unman thanks for the reply!
If my memory is serving me correctly, I first changed both the system storage in the template, then the private storage with the GUI arrows. I lowered them (after previously increasing them)
That’s all, so given it would have ignored it, I’m not sure what happen then.
But the errors above are what I received.
I will remove the “Suggestion to devs” part and modify my post to accurately reflect my lack of understanding of what went wrong.
If you have any other suggestions to how my post could be more helpful please tell me. Additionally I would be happy to mark your next response as a solution (it feels weird to mark my own non solution as a solution.)