Description of problem:
With the following partition setup in a kickstart file:
part /boot --fstype ext3 --size=50
part pv.1 --size=5800 --grow
volgroup rootvg pv.1
logvol / --vgname=rootvg --size=512 --name=root
logvol swap --fstype=swap --vgname=rootvg --recommended --name=swap
logvol /usr --vgname=rootvg --size=4000 --name=usr
logvol /var --vgname=rootvg --size=256 --name=var
logvol /export/data1 --vgname=rootvg --size=1 --grow --name=data1
I'm trying to install on a 10GB disk, but it is not growing pv.1 to fill the
rest of the disk. Instead I seem to be getting random sizes. In fact, as I
increase the --size parameter for pv.1, the partition it creates keeps getting
smaller. Actually, it appears that it is treating "size" as the amount of disk
space to keep free on the disk.
Are there any pre-existing non-linux partitions on the disk?
It doesn't seem to matter if there are or not. I've seen it on both of my
fc4test3 installs, one on a linux only disk (1 ext3 boot, 1 LVM) and one with
existing HFS+ partitions (on a mac).
On the linux only disk, can you see if clearpart --all makes any difference?
clearpart --all does not appear to make any difference.
Please try out the updates.img available from
http://people.redhat.com/clumens/158033-update/ and let me know if it makes any
difference. This is a test fix for a different problem that has some similar
behavior to what you're seeing.
Does not appear to have any effect (this is placed in the Fedora/base/ directory
for NFS installs, right?).
I'm unable to reproduce this on our later FC4 tests. Please test again with FC4
and let us know if it's still broken. I've seen similar symptoms regarding
using the size as the amount of free space to leave, and I believe it should be
Um, where do I get a later FC4 test?
Appears to be fixed in FC4 final