Red Hat Bugzilla – Bug 185272
anaconda doesn't handle all PE sizes in UI
Last modified: 2007-11-30 17:11:27 EST
Description of problem:
March 11th 2006 rawhide tree.
On my laptop I have an FC4 install on LVM. I created a new 15GB LV "rawhide" to
When I booted the installer, I chose "custom layout".
On the disk druid screen I could see my "VG0" volume group, my FC4 LVs and the
"rawhide" LV I created. I highlighted the "rawhide" LV and then clicked the Edit
button intending to assign "/" to it with the ext3 format.
However, when I clicked Edit, Anaconda threw up a trackback. I tried the "remote
host" option, but it seemed to hang, so I switched to tty2 and ran scp manually
to xfer the crash data.
Here first part of the crash data, I'll attach the whole crash data.
Traceback (most recent call last):
File "/usr/lib/anaconda/iw/partition_gui.py", line 1046, in editCb
File "/usr/lib/anaconda/iw/partition_gui.py", line 1126, in editLVMVolumeGroup
File "/usr/lib/anaconda/iw/lvm_dialog_gui.py", line 1001, in __init__
(self.lvmlist, sw) = self.createAllowedLvmPartitionsList(self.availlvmparts,
self.origvgrequest.physicalVolumes, self.partitions, origvgrequest.getPreExisting())
File "/usr/lib/anaconda/iw/lvm_dialog_gui.py", line 326, in
pesize = int(self.peCombo.get_active_value())
TypeError: int() argument must be a string or a number
Local variables in innermost frame:
sw: <gtk.ScrolledWindow object (GtkScrolledWindow) at 0xb7b2f504>
request: Existing Part Request -- mountpoint: None uniqueID: 4
type: physical volume (LVM) format: None badblocks: None
device: hda5 drive: hda primary: None
size: 128675.008301 grow: 0 maxsize: None
start: 29529423 end: 293055839 migrate: None fslabel: None origfstype:
physical volume (LVM)
bytesPerInode: 4096 options: 'None'
self: <lvm_dialog_gui.VolumeGroupEditor instance at 0xb736960c>
partlist: <WideCheckList object (GtkTreeView) at 0xb7b2f34c>
alllvmparts: [(4, 128675L, 1)]
part: (4, 128675L, 1)
partitions: <partitions.Partitions instance at 0xb73552ac>
store: <gtk.TreeStore object (GtkTreeStore) at 0xb7b2f324>
Created attachment 126012 [details]
The whole debug info generated by Anaconda
This crash is 100% reproducible. On another try, when I got to the disk druid
screen before clicking Edit and causing the crash, I went to tty2 and ran the
"lvm " command.
A lvdisplay showed all my LVs, but none where "available". Is that expected?
I ran vgchange and activated everything.
I went back to disk druid and clicked Edit. The crash still happened.
Is there anything "strange" abour your volume group? I'm not reproducing this
here in a quick test.
Can you run
lvm vgdisplay --options vg_name,vg_size,vg_extent_size
and show the output?
I don't think there is anything strange, other than the fact that in my FC4
install started out without LVM, but when I got a new laptop hard drive, I
booted to FC4 rescue mode and created the LVM by hand, then "cp -a" my FC4 data
back and then rebuilt my initrd.
So I migrated my FC4 to LVM (which has been working fine).
From inside my running FC4, here is the output of vgdisplay:
lvm> vgdisplay --options vg_name,vg_size,vg_extent_size
Incompatible options selected
--- Volume group ---
VG Name VG0
Metadata Areas 1
Metadata Sequence No 5
VG Access read/write
VG Status resizable
MAX LV 0
Cur LV 4
Open LV 3
Max PV 0
Cur PV 1
Act PV 1
VG Size 125.62 GB
PE Size 128.00 MB
Total PE 1005
Alloc PE / Size 656 / 82.00 GB
Free PE / Size 349 / 43.62 GB
VG UUID MTYLg4-n3IU-SrSW-576g-TIZ9-Wks2-C3obqC
Do you want this same output from tty2 during a rawhide install?
What's happening is that your PE size is 128M which is above what we generally
show in the UI.
For FC5 final, I've just bumped the "what we show" up to 128M. For FC6, we
should make this handle the case where someone does something we don't expect
and just add that value to the list.