The __init__ constructor of the Device class in the module GLIStorageDevice class will fail if a disk doesnt contain a recognisable label. self._parted_disk = parted.PedDisk.new(self._parted_dev) --> this part will fail There should be a check of the disk contains a label, if not something like disk = device.disk_new_fresh(parted.disk_type_get(label)) disk.commit() should be executed. Failure of parted.PedDisk.new(self._parted_dev) gives the following error >>> dev = parted.PedDevice.get('/dev/sdb') >>> disk = parted.PedDisk.new(dev) Traceback (most recent call last): File "<stdin>", line 1, in ? parted.error: Error: Unable to open /dev/sdb - unrecognised disk label. Output of fdisk jabba root # fdisk -l /dev/sdb Disk /dev/sdb: 4294 MB, 4294967296 bytes 255 heads, 63 sectors/track, 522 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk /dev/sdb doesn't contain a valid partition table
As a side note it seems that the install.log did not get anything written to it. We need to allow for it to gracefully fail.
*** Bug 93114 has been marked as a duplicate of this bug. ***
This is a known issue. I've been meaning to fix it, but it's more complex than just creating a new disklabel (different arches and all). The reason that you have an empty install.log is because it failed in the FE before the install started.
I've added some code so that the frontend doesn't fail: self._parted_dev = parted.PedDevice.get(self._device) try: self._parted_disk = parted.PedDisk.new(self._parted_dev) except: self._parted_disk = self._parted_dev.disk_new_fresh(parted.disk_type_get(archinfo[self._arch]['disklabel']))
test...ignore
Even though I haven't implemented the backend part of this yet, the installer doesn't work on other architectures yet, so it isn't currently an issue. Closing.
First of all, my apologies for my late respons ... I tested the snapshot installer-20050613.tar.bz2. I encounterd the same problem but in the x86ArchitectureTemplate class. At the same time i noticed that the installer also tries to format the 'free' space on the disk. I wrote a quick patch (and i tested this) that solves both issues. brgds.
Created attachment 61305 [details, diff] patch against x86ArchitectureTemplate.py patch against x86ArchitectureTemplate.py
(In reply to comment #8) > Created an attachment (id=61305) [edit] > patch against x86ArchitectureTemplate.py > > > patch against x86ArchitectureTemplate.pyGLI: Also the contents of /var/log/install.log June 15 2005 19:30:43 - partitioning: Processing /dev/sdb... GLI: June 15 2005 19:30:43 - Partitioning: Second pass... GLI: June 15 2005 19:30:43 - Partitioning: Third pass....creating partitions GLI: June 15 2005 19:30:43 - Drive has 8388608 sectors GLI: June 15 2005 19:30:43 - Partition 1 has 1027MB GLI: June 15 2005 19:30:43 - Adding primary partition 1 from 0 to 2103296 GLI: June 15 2005 19:30:43 - Formatting partition 1 as ext3 with: mkfs.ext3 /dev/sdb1 GLI: June 15 2005 19:30:44 - Partition 1.03125 has 3069MB GLI: June 15 2005 19:30:44 - Exception received during 'Partition': PartitionFormatError :FATAL: partition: Unknown partition type free GLI: June 15 2005 19:30:44 - Traceback (most recent call last): GLI: June 15 2005 19:30:44 - File "../../GLIClientController.py", line 151, in run self._install_steps[self._install_step][0]() GLI: June 15 2005 19:30:44 - File "../../templates/x86ArchitectureTemplate.py", line 335, in partition raise GLIException("PartitionFormatError", 'fatal', 'partition',"Unknown partition type "+newpart['type']) GLI: June 15 2005 19:30:44 - GLIException: PartitionFormatError :FATAL: partition: Unknown partition type free
A slightly modified version of that patch has been applied. I changed it to look at tmp_parts_new[device].get_disklabel() first before going to the default disklabel type for that particular arch. Thanks.
Moving to Release Media/Installer.
