Inconsistent output for "snap refresh"/"snap refresh <specific_snap>"
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| snapd |
Triaged
|
Medium
|
Unassigned | ||
Bug Description
Using "snap set core experimental.
Example: With snap "lxd" installed and running containers, snapd would previously shut down those and possibly restart them afterwards. Now, (tested with snapd 2.50 on Ubuntu 20.04LTS, arm64/amd64 architectures) what you get to see is
* "snap refresh" --> "All snaps up to date."
* "snap refresh lxd" --> "error: cannot refresh "lxd": snap "lxd" has running apps (lxc)"
While the current behaviour is IMHO very much in line with the feature definition ("ensure the least amount of disruption to your work") *now*, the output of both commands above should be identical (e.g. always showing errors for "non-refreshable" snaps) in said situation.
CVE References
| Changed in snapd: | |
| importance: | Undecided → High |
| status: | New → Triaged |
| Changed in snapd: | |
| importance: | High → Medium |
| Harm van Bakel (hvbakel) wrote | #2 |
This is a security issue!
I just had a vuln. package on my system. fixed version already in store. snap telling me "All snaps up to date".
So I can never be sure, my system is up to date until I go through all my snaps by hand.