Description of problem:
At this point I don't know if this is a problem in anaconda, parted, or the
kernel. An install of Fedora failed due to not finding the partition that it
had just created.
I was installing to /dev/cciss/c0d0, it should have created c0d0p0 p1 p3 and p3.
The install failed that it could not find the partitions. I dropped to a shell
and confirmed that the /dev/cciss/c0d0 (aka the raw disk) was the only dev file.
After a reboot however I _did_ see that the /dev/cciss/c0d0p* dev files did exist.
So, whatever is supposed to make these devices after parted is run failed. I
don't know how that works under anaconda, does it use udev?
I will try this on an x86 system but I am guessing that installing to any disk
where the partitions did not exist when the installer booted will fail.
It is difficult to tell when this failed since normally my installs are to a
system that already has the partitions created.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. make sure you have a disk with no partitions on it
2. install to that disk
3. anaconda fails
Created attachment 303628 [details]
dump file from anaconda
Created attachment 303629 [details]
anaconda.log from failed install
I just did an install (to a disk that had pre-existing partitions) and from the
installed system I partitioned a disk that had no pre-existing partitions with
parted. The /dev/ entries do get created in this case so I think that rules out
the problem being in the kernel, udev or parted.
My guess is whatever anaconda does to make sure new /dev/ entries get created
isn't working. This was on ia64, I will do an test on my x86_64 system as soon
as I get in the office to see if this is arch specific.
*** This bug has been marked as a duplicate of 439633 ***