journal contains errors on each service restart "keyword is not comprised only of uppercase ASCII letters or digits"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
libvirt (Ubuntu) |
Fix Released
|
Low
|
Unassigned | ||
Jammy |
Won't Fix
|
Low
|
Unassigned | ||
Kinetic |
Fix Released
|
Low
|
Unassigned |
Bug Description
Some systems have been reported to emit errors to the log on each libvirtd restart.
An example is:
libvirtd[15231]: internal error: The keyword is not comprised only of uppercase ASCII letters or digits
libvirtd[15231]: internal error: A field data length violates the resource length boundary.
...
Such messages can almost always safely be ignored.
It almost always points to device/firmware reporting invalid VPD data.
Libvirt is ok to stumble over this, but maybe reporting it as an error is too much.
Note:
If you are affected and you wonder which device is causing this you might run the following:
$ for i in $(sudo find /sys -name '*vpd' | awk -F '/' '{gsub(":","_"); gsub("\\.","_"); print $6}'); do virsh nodedev-dumpxml pci_${i} | grep -q "capability type='vpd'" || echo "device ${i} has invalid VPD data"; done
device 0000_02_00_2 has invalid VPD data
device 0000_02_00_0 has invalid VPD data
device 0000_02_00_3 has invalid VPD data
device 0000_02_00_1 has invalid VPD data
So in my case I've found 4 entries being broken, and when looking into those they are 4 functions of the same device which is the one with the bad firmware.
If you can check if updating the firmware fixes the issue for good.
Related branches
- git-ubuntu bot: Approve
- Sergio Durigan Junior (community): Approve
- Canonical Server Reporter: Pending requested
-
Diff: 278 lines (+250/-0)4 files modifieddebian/changelog (+9/-0)
debian/patches/series (+2/-0)
debian/patches/ubuntu/lp-1990499-virt-aa-helper-allow-common-riscv64-loader-paths.patch (+58/-0)
debian/patches/ubuntu/lp-1990949-virpcivpd-reduce-errors-in-log-due-to-invalid-VPD.patch (+181/-0)
Changed in libvirt (Ubuntu): | |
status: | New → Confirmed |
importance: | Undecided → Low |
While inconvenient the importance is not very high.
I think we want to fix it upstream and in the future, but not backport it to active releases.
But I'm happy to be convinced otherwise ...
I'm starting to suggest a fix to upstream anyway to have a discussion on it. /listman. redhat. com/archives/ libvir- list/2022- September/ 234517. html
=> https:/