Description of problem: When writing a Fedora Core ISO image to CD, cdrecord is invoked with the -nopad option. This seems to universally break the anaconda mediacheck option. As a result, the written CD cannot be checked. I am not sure whether this is a bug with a) nautilus-cd-burner, b) the way Fedora Core ISO images are MD5SUM-embedded c) the way anaconda checks the images against the sum Version-Release number of selected component (if applicable): nautilus-cd-burner-2.10.0-2 How reproducible: 100% Steps to Reproduce: 1. Write a Fedora Core 4 disc1 CD using nautilus-cd-burner 2. Boot the CD with 'linux mediacheck nocddma' Actual results: FAIL Expected results: PASS
I should mention that there is no method of checking CDs written in this way. MD5SUMs of ISO images before writing do not match MD5SUMs of CDs unless they are written with -pad.
I think I've just run into this same problem: I burned a set of FC4 CDs using nautilus (right click on ISO file, choose write to cd), and they were all bad. I tested with dd if=/dev/cdrom bs=512 | cmp - FC-xxx.iso and in all cases the dd came up with "Input/output error" a bit before the end of the CD (anywhere from 2K to several hundred K before the end). The cmp didn't show any differences before that, so I don't think this is the CD burner's fault. What's even more disturbing is that anaconda's media check facility claimed that two of the five CDs were good, when the dd test proves they aren't. Perhaps I should file that as a separate bug? This was with FC4 tip, ie nautilus-cd-burner-2.10.0-2 cdrecord-2.01.1-9.0.FC4.1 and whatever version of anaconda is on the FC4 release CDs. The hardware is a Dell 470n that just arrived yesterday ...
Seems to be fixed upstream now: http://bugzilla.gnome.org/show_bug.cgi?id=315068 Closing rawhide, since this will be in fc5.
*** Bug 139231 has been marked as a duplicate of this bug. ***