The following was filed automatically by anaconda: anaconda 11.5.0.47 exception report Traceback (most recent call first): File "/usr/lib/anaconda/storage/devicelibs/swap.py", line 46, in mkswap raise SwapError("mkswap failed for '%s'" % device) File "/usr/lib/anaconda/storage/formats/swap.py", line 142, in create swap.mkswap(self.device, label=self.label) File "/usr/lib/anaconda/storage/deviceaction.py", line 273, in execute options=self.device.formatArgs) File "/usr/lib/anaconda/storage/devicetree.py", line 659, in processActions action.execute(intf=self.intf) File "/usr/lib/anaconda/storage/__init__.py", line 234, in doIt self.devicetree.processActions() File "/usr/lib/anaconda/packages.py", line 117, in turnOnFilesystems anaconda.id.storage.doIt() SwapError: mkswap failed for '/dev/vda2'
Created attachment 340854 [details] Attached traceback automatically from anaconda.
This happened while trying to do an install of rawhide into a KVM virtual machine. The following is an approximation of the layout I chose: /dev/vda1 /boot 128MB /dev/vda2 swap 1024MB /dev/vda3 / the rest This was on a 10G qcow2 disk image that was initialized at the beginning of the Anaconda session. /boot was ext3 and / was btrfs. It appears the crash happened during mkswap.
I tried again, letting it create the default layout (though I changed / from ext4 to btrfs). It still crashes, but this time with pvcreate for /dev/sda2. I'm going to try once more, simply letting it do the default layout with no changes.
Still crashes, so it definitely had nothing to do with btrfs. And that should say /dev/vda2.
Whatever kind of device /dev/vda is, you are experiencing a rather serious problem with it. There are tons of messages like these in your error log: <3>end_request: I/O error, dev vda, sector 97 <3>Buffer I/O error on device vda1, logical block 12 <4>lost page write due to I/O error on vda1 <3>end_request: I/O error, dev vda, sector 105 <3>end_request: I/O error, dev vda, sector 113 <3>end_request: I/O error, dev vda, sector 121 <3>end_request: I/O error, dev vda, sector 145
It's just a qcow2 file sitting on my Fedora 10 partition (ext3). I've used the same to install rawhide before and it worked (maybe a week or so ago). Here's the command I used to create the qcow2 file: qemu-img create -f qcow2 gidux-kvm-rawhide-kde.qcow2 10G Do you think those errors could be indicative of an issue I should investigate with my hard drive? Could a KVM virtual machine see that far into the host, so to speak? Palimpset says my laptop drive is okay, and I haven't been having issues. Otherwise, do you think there's an issue with qcow2 handling?
Do you have any thoughts on this? As I said before, I'm not accessing a physical device, but a file. Would it still see errors related to the physical device that the file is on, or could this be an indication of a bug in the qcow2 driver? Once F11 final is out, I plan on doing a fresh install and formatting my main partition with ext4, so I'll be able to try this again with the qcow2 file on an ext4 / partition instead of ext3 (though, I would think we'd want the same behavior regardless of the filesystem).
The installer does not crash if my layout does not include a swap partition.
I know this won't necessarily help identify the problem, but I wanted to note that I am currently installing Fedora 11 Preview in VirtualBox 2.2.4 OSE (from RPM Fusion), with the default partition layout, and Anaconda did not error out. The virtual disk is on my ext3 partition.
I just tried once more, and it worked this time! I noticed there were updates to libvirt and virt-manager today, so I don't know if those are related, but it appears to be working now. I'm counting this closed.
Created attachment 347584 [details] Attached traceback automatically from anaconda.