ext-tests/ext Fix var-mount/scsi-id failure on rawhide #2943
Conversation
|
@dustymabe you were so quick to review I could not post my follow-up comment :) Let's hold this for now : coreos/coreos-assembler#3772 (comment) EDIT: now it's fixed :) |
Update the scsci-id test to set a WWN for the disk, and use reliable udev symlinks to adjust for a change in sg3_utils [1] Note that the `wwn` value set is converted to base 16 by QEMU, so the symlink in the ignition config must reflects it. This requires coreos/coreos-assembler#3772 See coreos/fedora-coreos-tracker#1670 [1] https://listman.redhat.com/archives/dm-devel/2023-March/053645.html
jbtrystram
force-pushed
the
fix/mount.scsi-id
branch
from
5f490af
to
d93cc53
Compare
That value is one we add so that we can ID the related block device inside the VM. My understanding is that in this case the driver takes the decimal value and converts it to hexadecimal underneath, so the symlink that shows up in the VM is hexadecimal. |
|
@c4rt0 looks at coreos/coreos-assembler#3772 for how the |
An update in rawhide broke /dev/disk-by symlinks that the scsi-id relied on. [1] fixed the test but i forgot to re-enable it [1] coreos#2943
An update in rawhide broke /dev/disk-by symlinks that the scsi-id relied on. [1] fixed the test but i forgot to re-enable it [1] #2943
Update the scsci-id test to set a WWN for the disk, and use reliable udev symlinks to adjust for a change in sg3_utils [1]
Note that the
wwnvalue set is converted to base 16 by QEMU, so the symlink in the ignition config must reflects it.This requires coreos/coreos-assembler#3772 See coreos/fedora-coreos-tracker#1670
[1] https://listman.redhat.com/archives/dm-devel/2023-March/053645.html