Description of problem:
This is branching from bug 616273 comment 98. Please see this bug for the
"'NoneType' object has no attribute 'length'' exception of anaconda when trying to install to existing partitions on multipath device
The same phenomenon has occurred with different operations / conditions.
See the following comment for details:
The problem was last confirmed with RHEL6 Snapshot10 with the bz616273.img
mentioned in this comment:
Version-Release number of selected component (if applicable):
RHEL 6 Snapshot 10 + bz616273.img
every time (?)
Steps to Reproduce:
1. Prepare 1 multipath device
2. Create partitions to install RHEL6 from rescue mode
3. Reboot and start RHEL6 installation
4. Select "Specialized Storage Device"
5. Select the multipath device in device selection
6. Select "Create Custom Layout"
7. Put "/" and "/boot" on the existing partitions created in the step 2
8. Click "Next"
anaconda crashes with the error "'NoneType' object has no attribute 'length'"
installation proceeds normally
17:04:27,667 CRITICAL: anaconda 13.21.68 exception report
Traceback (most recent call first):
File "/tmp/updates/iw/partition_gui.py", line 383, in _createStripe
File "/tmp/updates/iw/partition_gui.py", line 302, in setDisplayed
self.stripe = self._createStripe(obj)
File "/tmp/updates/iw/partition_gui.py", line 356, in __init__
File "/tmp/updates/iw/partition_gui.py", line 1254, in treeSelectCB
dcCB = self.barviewActivateCB)
AttributeError: 'NoneType' object has no attribute 'length'
Created attachment 439288 [details]
Debug patch for anaconda to treat singlepath SCSI devices as multipath device
The problem can be reproduced without real multipath hardware
using the attached patch.
can you please try to retest with
in the kernel command line?
The patch ensures multipath parted objects are not torn down when they shouldn't.
Please zip and attach all log files in case of another traceback.
With bz624175.img, I got "Invalid Device Specification" as in this BZ:
Logs uploaded in issuetracker.
Scratch comment 7.
Since the patch from Comment 4 helped get past the point of error, I'm marking this as a dup of 624175.
I'll be tracking progress on the error Jun'ichi is reporting in Comment 5 in bug 624497.
*** This bug has been marked as a duplicate of bug 624175 ***