The Snap problems are back!

trenchanter

The saga continues! In the last episode I fixed my system and removed/reinstalled snapd and installed a couple of snaps which appeared to be working just fine!

But lo, it was not to be. On a restart my dreams were shattered - for a reason I can't determine, snapd.service is not starting on boot, and can't be started manually with sudo systemctl start snapd.service.

I've scoured the Snapcraft forums but haven't found a solution - I've also made a post there, but it is likely to remained untouched until the 6th, so I'll post the same information here in the hopes a local can help me out.

Cheers!

Output from: snap version
snap 2.39.3
snapd unavailable
series -

Output from: snap find
trenchanter@solus ~ $ snap find error: cannot list snaps: cannot communicate with server: Get http://localhost/v2/find?scope=wide&section=featured: dial unix /run/snapd.socket: connect: connection refused

Output from: systemctl status snapd.service
snapd.service - Snappy daemon Loaded: loaded (/usr/lib/systemd/system/snapd.service; disabled; vendor pres> Active: failed (Result: exit-code) since Tue 2019-12-31 10:33:20 NZDT; 25min> Process: 3521 ExecStart=/usr/lib64/snapd/snapd (code=exited, status=1/FAILURE) Main PID: 3521 (code=exited, status=1/FAILURE)

Output from: sudo SNAPD_DEBUG=1 /usr/lib/snapd/snapd
AppArmor status: apparmor is enabled and all features are available 2019/12/31 11:13:47.018042 cmd_linux.go:70: DEBUG: re-exec not supported on distro "solus" yet 2019/12/31 11:13:47.022514 backend.go:115: snapd enabled NFS support, additional implicit network permissions granted cannot run daemon: cannot reload snap-confine apparmor profile: cannot load apparmor profiles: exit status 2 apparmor_parser output: File /etc/apparmor.d/usr.lib.snapd.snap-confine not found, skipping...

Output from: journalctl -u snapd --no-pager
-- No entries --

Output from: sudo journalctl -xe (< I've put this in pastebin)

Contents of: /var/lib/snapd/state.json
{"data":{"patch-level":6,"patch-sublevel":1,"refresh-privacy-key":"REDACTED"},"changes":{},"tasks":{},"last-change-id":0,"last-task-id":0,"last-lane-id":0}

DataDrake

Make sure the socket is enabled:

sudo systemctl enable snapd.socket
sudo systemctl status snapd.socket

Do a full usysconf for good measure:

sudo usysconf run -f

Try again with the correct path this time:

sudo SNAPD_DEBUG=1 /usr/lib64/snapd/snapd

The 64 is important as the file it should be looking for is /etc/apparmor.d/usr.lib64.snapd.snap-confine.

trenchanter

Hi DataDrake
I enabled snapd.socket but on boot the status would always "Failed" but could be started manually with systemctl start snapd.socket. I then tried sudo usysconf run -f with no errors.

I then tried the following command to see if this had affected snapd.service at all:
sudo systemctl status snapd.service

This caused the system to enter an unrecoverable hang and boot to userspace no longer happens.
By using Ctrl + Alt + F2 I ran a check of installed packages and discovered the kernel is corrupted.

Checking integrity of linux-current Broken Access denied: /usr/lib/kernel/initrd-com.solus-project.current.5.3.18-140 Corrupted file: /lib/modules/5.3/18-140.current/modules.builtin.bin Corrupted file: /lib/modules/5.3/18-140.current/modules.devname Corrupted file: /lib/modules/5.3/18-140.current/modules.softdep

I've tried various install/reinstall/upgrade commands but always run into the same problem.

Program terminated. Could not fetch destination file "https://mirrors.rit.edu/solus/packages/shannon/eopkg-index.xml.xz.sha1sum": [Errno 14] curl#6 - "Could not resolve host: mirrors.rit.edu"

I hoped I could load another version of the kernel, but sudo clr-boot-manager list-kernels provides the following:
[FATAL] cbm (../srg/bootloaders/systemd-class.c:l507): sd_class_get_default_kernel Unable to parse loader config com.solus-projec.current.5.3.18-140

The output of sudo clr-boot-manager update:
[FATAL] cbm (../src/bootman/update.c:L154: FATAL: Cannot mount boot device /dev/sda1 on /boot: No such device

And once again I'm stumped - please send help!

trenchanter

trenchanter
I've managed to get the system back up and running on the LTS Kernel, I've theoretically reinstalled the current kernel with eopkg - and eopkg check indicates it's fine - but clr-boot-manager will not allow me to select it for next boot.

Kernel snafu aside, the behaviour of snapd.socket and snapd.service remain the same - neither start on reboot of the machine. Socket can be started, but here is the output of sudo systemctl start snapd.service:

snapd.service - Snappy daemon
Loaded: loaded (/usr/lib/systemd/system/snapd.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Fri 2020-01-03 16:09:00 NZDT; 9s ago
Process: 2727 ExecStart=/usr/lib64/snapd/snapd (code=exited, status=1/FAILURE)
Main PID: 2727 (code=exited, status=1/FAILURE)

Jan 03 16:09:00 solus systemd[1]: snapd.service: Service RestartSec=100ms expired, scheduling restart.
Jan 03 16:09:00 solus systemd[1]: snapd.service: Scheduled restart job, restart counter is at 5.
Jan 03 16:09:00 solus systemd[1]: Stopped Snappy daemon.
Jan 03 16:09:00 solus systemd[1]: snapd.service: Start request repeated too quickly.
Jan 03 16:09:00 solus systemd[1]: snapd.service: Failed with result 'exit-code'.
Jan 03 16:09:00 solus systemd[1]: Failed to start Snappy daemon.
Jan 03 16:09:00 solus systemd[1]: snapd.service: Triggering OnFailure= dependencies.

After this happens. I check systemctl status snapd.socket and see that it has now moved back to a failed state:

● snapd.socket - Socket activation for snappy daemon
   Loaded: loaded (/usr/lib/systemd/system/snapd.socket; enabled; vendor preset: enabled)
   Active: failed (Result: service-start-limit-hit) since Fri 2020-01-03 16:09:00 NZDT; 3min 44s ago
   Listen: /run/snapd.socket (Stream)
           /run/snapd-snap.socket (Stream)