Manual partitioning during installation crashes installer

I was trying to do a manual partition during install to set up root as RAID1, but the installer keeps crashing when I try to set the options for a partition.

The standard Centos 7 minimal install is working fine.

Traceback (most recent call first):
  File “/usr/lib64/python2.7/site-packages/gi/types.py”, line 113, in function
    return info.invoke(*args, **kwargs)
  File “/usr/lib64/python2.7/site-packages/pyanaconda/ui/gui/spokes/custom.py”, line 1896, in _populate_right_side
    self._typeCombo.set_active(type_index_map[device_type])
  File “/usr/lib64/python2.7/site-packages/pyanaconda/ui/gui/spokes/custom.py”, line 2614, in on_selector_clicked
    self._populate_right_side(selector)
  File “/usr/lib64/python2.7/site-packages/pyanaconda/ui/gui/spokes/custom.py”, line 2258, in _show_mountpoint
    self.on_selector_clicked(page.members[0])
  File “/usr/lib64/python2.7/site-packages/pyanaconda/ui/gui/spokes/custom.py”, line 2633, in on_page_clicked
    self._show_mountpoint(page=page, mountpoint=mountpointToShow)
  File “/usr/lib64/python2.7/site-packages/pyanaconda/ui/gui/spokes/lib/accordion.py”, line 146, in _onExpanded
    cb(obj.get_child())
TypeError: Argument 1 does not allow None as a value

Thanks for the traceback. It’s very helpful.

Can you share the complete partition table you’d like to create? You would like to setup a BTRFS RAID1 using two whole disk drives for root filesystem, correct? How about /boot which I believe needs to be ext4?

I am able to reproduce the problem, btw. One thing that confuses the installer right now is if the underlying disks have partioning/data on them. I was able to go farther by going into the console(ALT+F2) and wipe the drives with wipefs -a sdX command and then proceed to partition. You might want to experiment this approach. But honestly, I am unsure as to what a sensible RAID1 partitioning scheme looks like given there is still a need for a ext4 /boot partition. Please do share your thoughts.

I am running into the installer bug with completely blank disks.

I was actually trying to sort out a disk corruption issue and thought that RAID1 might provide a bit of resiliency, rather than just doing a rebuild whenever scrub comes up with an error. I think I have now traced it to an issue with VirtualBox.

You make a good point about the boot partition, although that could be cloned onto both disks. Also it is easier to recover from a trashed boot partition than root partition. As I am using a VM, I can just create /boot on a dedicated ‘disk’.

Regardless, I have managed to get the system installed on RAID1, but cannot get it to boot properly. Either it cannot mount root and drops me to emergency mode where I can manually mount root and continue, or it kernel panics.

Trying to install the latest version as we speak, getting the crash. I can avoid the bug by reclaiming disk space and allowing for automatic partitioning

Reclaiming space in the destination selection menu that is

I think this issue is definitely fixed in the current iso, version 3.8-0. Just want to close the loop on this topic.