"vg0 is not a valid name for a volume group" if VG already exists
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| subiquity (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned | ||
Bug Description
I'm installing Ubuntu 18.04.2 amd64 server on a BIOS booted VirtualBox VM. It has two storages, on which I previously installed Ubuntu using the same installer. Then, I created an MD RAID-1 backed by a total of two component devices, which are partitions, one on each of the two storages. On top of this RAID array, I had the installer create an LV PV and VG.
I am now re-installing using the same disks. Subiquity's manual partitoning screen starts with a display of two blank HDDs - it states that there is nothing on them, no partitions, no other block devices.
I start partitioning this:
HDD1 (10GB)
PT1: 1MB, bios_grub
PT2: 1GB, ext4, /boot
PT3: 8.997GB, comp. of md0
HDD2 (10GB)
PT1: 1MB, unused
PT2: 1GB, unused
PT3: 8.997GB, comp. of md0
MD
MD0: 8.989GB, RAID-1
Then I try to make md0 a PV for a VG named "vg0" (pre-filled default name). At this point, the installer reports: "vg0 is not a valid name for a volume group" (screenshot attached).
Which is nonsense unless you know that a vg0 already exists on these drives - but the installer doesn't tell you, rather claims there was nothing there.
Naming the VG "vg1" instead succeeds and the installation continues.
| Michael Hudson-Doyle (mwhudson) wrote | #1 |
| Launchpad Janitor (janitor) wrote | #2 |
| Changed in subiquity (Ubuntu): | |
| status: | New → Confirmed |
Yeah, I've hit this too. As with everything else, the "reusing existing partitions" work is the answer, because that will include probing for existing LVM/mdraid/etc devices.