Description of problem:
Install os on a big image(2T), screen shows garbage during format disk.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Create a big image.
#qemu-img create -f qcow2 big.img 2T
2.Start qemu-kvm with following command line.
-M pc \
-cpu SandyBridge \
-m 2G \
-smp 4,sockets=2,cores=2,threads=1 \
-name rhel6 \
-uuid 990ea161-6b67-47b2-b803-19fb01d30d12 \
-smbios type=1,manufacturer='Red Hat',product='RHEV Hypervisor',version=el6,serial=koTUXQrb,uuid=feebc8fd-f8b0-4e75-abc3-e63fcdb67170 \
-k en-us \
-rtc base=localtime,driftfix=slew \
-monitor stdio \
-qmp tcp:0:6779,server,nowait \
-boot menu=on \
-bios /usr/share/seabios/bios.bin \
-serial unix:/tmp/console0,server,nowait \
-spice port=5900,disable-ticketing \
-vga qxl \
-drive file=/home/big.img,if=none,id=drive-virtio-disk0,format=qcow2,cache=none,werror=stop,rerror=stop,aio=threads \
-device virtio-blk-pci,scsi=off,bus=pci.0,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1,serial="QEMU-disk0" \
-usb -device usb-tablet,id=input0 \
-drive file=/home/boot.iso,if=none,media=cdrom,id=drive-ide0,readonly=on,format=raw \
-device ide-drive,bus=ide.0,unit=0,drive=drive-ide0,id=ide0,bootindex=0 \
-netdev tap,id=tap1,vhost=on \
-device virtio-net-pci,netdev=tap1,id=nic1,mac=52:54:00:ae:f9:55 \
Screen shows garbage during format disk(format disk will take a long time).
Use vga std not hit this problem.
Created attachment 998637 [details]
Happens on RHEL-7 host too.
Looks like a guest bug.
Assigning to xorg qxl driver for investigation.
What is boot.iso? If it's the RHEL6 installer, is it using the QXL driver? (try to get the Xorg log during install)
(if not, it looks like a VGA emulation bug, and should be reassigned to qemu)
Happens with a rhel 6.6 server iso, and the installer is using QXL. Happens when big.img grows to about 5GB here. Could not get useful logs at this point.
After managing to get an ssh connection to the VM (add 'sshd' to the boot options, and then enable eth0 in the graphical installer), I've managed to get
[ 722.011] Out of memory allocating 31460 bytes
[ 722.011] Out of mem - stats
[ 722.015] max system bytes = 50319360
[ 722.015] system bytes = 50319360
[ 722.015] in use bytes = 48957432
http://lists.freedesktop.org/archives/spice-devel/2015-March/019229.html fixes the issue for me.
reproduced according to instructions. The disk size rose to 32 GB in my case however.
If you copied the fixed QXL rpm to the iso in the Packages dir, I think you changed the QXL driver which would be used after the installation completes. the driver which is used during the installation must be embedded uncompressed in some .img file or such which is uncompressed/loopback mounted before the installation starts. Ie not that trivial to upgrade the installer.
What I did was that I started from a RHEL6.6/6.7 system, upgraded the QXL driver to the latest version, yum install'ed anaconda (https://brewweb.devel.redhat.com/packageinfo?packageID=98 ), and then opened a root terminal, and started anaconda from there. So quite different from booting the installer directly from the ISO, but much easier. If you want to try an installer, there are probably some 6.7 installer isos with the right driver (?)
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.