Created attachment 594503 [details]
first 34 sectors of USB drive
I blasted a custom Fedora x86 DVD image onto a cheap USB key drive, and then attempted to install from it. The installation halts during storage probing. After some digging, I discovered the following reproducer outside of anaconda with pyparted:
In : import _ped
In : xxx = _ped.device_get("/dev/sdb")
In : _ped.Disk(xxx)
However, this is reproducible outside of pyparted by just attempting "parted /dev/sdb p". A core dump gives the following:
#0 0x0000003f90c362a5 in raise () from /lib64/libc.so.6
#1 0x0000003f90c37bbb in abort () from /lib64/libc.so.6
#2 0x0000003f90c75ffe in __libc_message () from /lib64/libc.so.6
#3 0x0000003f90c7c606 in malloc_printerr () from /lib64/libc.so.6
#4 0x00007fe666f592fa in mac_read (disk=0x1eedc10) at mac.c:832
#5 0x00007fe666f44200 in ped_disk_new (dev=0x1f044b0) at disk.c:204
#6 0x00007fe66718ba2a in _ped_Disk_init (self=0x1f68f38, args=<optimized out>, kwds=<optimized out>) at src/pydisk.c:425
(irrelevant python stuff omitted)
This is funny, because there shouldn't be mac_anything involved. I am attaching the first 34 sectors of the device in question.
Maybe try to /dev/zero the drive and verify its image integrity? I was playing with partition tables just to find out, obviously, handling all kinds of corruption is not a duty of parted.
*** Bug 854168 has been marked as a duplicate of this bug. ***
It ends up that the problem was in max_read, it would change the sector_size to 2048 without re-allocating the read buffer.
parted-3.1-7.fc18 has been submitted as an update for Fedora 18.
You can test this with a livecd by downloading it and installing it before running liveinst.
I have updated parted and liveinst now can be run on TC5.
Discussed at 2012-09-05 blocker review meeting. Accepted as a blocker per criterion "The installer must boot (if appropriate) and run on all primary architectures, with all system firmware types that are common on those architectures, from default live image, DVD, and boot.iso install media when written to an optical disc and when written to a USB stick with at least one of the officially supported methods" - it appears this bug is pretty likely to occur on any USB installation. Three different people have hit it with multiple sticks, so far.
This is fixed in Alpha TC6 (anaconda 18.6.5).
Ah, bad bad me. This is about parted update, and it's not still in stable updates yet. Reverting to VERIFIED.
parted-3.1-7.fc18 has been pushed to the Fedora 18 stable repository. If problems still persist, please make note of it in this bug report.