Bug in /etc/update-motd.d/50-landscape-sysinfo script - incorrect variable used for load threshold calculation
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
landscape-client (Ubuntu) |
In Progress
|
Undecided
|
Spencer Runde |
Bug Description
[ Impact ]
* The CPU load is incorrectly calculated, and users won't get system info in
the message-of-the-day (motd).
* The `cores` variable was misnamed; it should be `CORES` since bash variables are case-sensitive
* Affects landscape-client on noble, mantic, jammy, and focal.
[ Test Plan ]
Steps to Reproduce:
1. Open the /etc/update-
2. Notice the line where THRESHOLD is set: THRESHOLD=
3. The correct variable should be CORES instead of cores.
To trigger MOTD and view incorrect calculation:
sudo run-parts /etc/update-motd.d/
Notice message: " System information disabled due to load higher than 1.0" even on systems with multiple cores.
[ Where problems could occur ]
* This fix will necessarily run the system process info on more systems during MOTD. If the CPU load threshold was poorly chosen, then overly-loaded systems might be slow during MOTD while landscape-sysinfo runs
Related branches
- Ubuntu Sponsors: Pending requested
- git-ubuntu import: Pending requested
-
Diff: 28 lines (+8/-1)2 files modifieddebian/changelog (+7/-0)
debian/landscape-sysinfo.wrapper (+1/-1)
- Ubuntu Sponsors: Pending requested
- git-ubuntu import: Pending requested
-
Diff: 28 lines (+8/-1)2 files modifieddebian/changelog (+7/-0)
debian/landscape-sysinfo.wrapper (+1/-1)
- Ubuntu Sponsors: Pending requested
- git-ubuntu import: Pending requested
-
Diff: 28 lines (+8/-1)2 files modifieddebian/changelog (+7/-0)
debian/landscape-sysinfo.wrapper (+1/-1)
- Ubuntu Sponsors: Pending requested
- git-ubuntu import: Pending requested
-
Diff: 28 lines (+8/-1)2 files modifieddebian/changelog (+7/-0)
debian/landscape-sysinfo.wrapper (+1/-1)
- git-ubuntu import: Pending requested
-
Diff: 28 lines (+8/-1)2 files modifieddebian/changelog (+7/-0)
debian/landscape-sysinfo.wrapper (+1/-1)
Hi Markus,
Thank you for your bug report. We have an SRU in progress for this, and it should land shortly.