I thought this could be a regression introduced during the python
migration, but I've just reproduced the error on a pre-python-migration
branch (r166), so it doesn't look like it.
This is kind of expected, though, as l-m-c will use parted/sfdisk to
partition your disk unless you pass --no-part to it. And I don't
remember seeing any code in there to umount stuff before partitioning.
(FWIW, l-m-c raises a similar error if you use --no-part as it will try
to run mkfs on a mounted partition)
This is a trace of the failure:
+ for func in '$DEPLOY_STEPS'
+ create_partitions
+ '[' /dev/sdb ']'
+ sudo parted -s /dev/sdb mklabel msdos
Error: Partition(s) on /dev/sdb are being used.
I thought this could be a regression introduced during the python migration
migration, but I've just reproduced the error on a pre-python-
branch (r166), so it doesn't look like it.
This is kind of expected, though, as l-m-c will use parted/sfdisk to
partition your disk unless you pass --no-part to it. And I don't
remember seeing any code in there to umount stuff before partitioning.
(FWIW, l-m-c raises a similar error if you use --no-part as it will try
to run mkfs on a mounted partition)
This is a trace of the failure:
+ for func in '$DEPLOY_STEPS'
+ create_partitions
+ '[' /dev/sdb ']'
+ sudo parted -s /dev/sdb mklabel msdos
Error: Partition(s) on /dev/sdb are being used.
status triaged
importance high