Bugzilla – Bug 1057449
[2.9.0] Qemu freezes when installing SLE15
Last modified: 2017-09-08 19:47:03 UTC
## observation When trying to install SLES 15, qemu hangs and stops responding. No errors are in the logs. Appears only if perform installation with X11, in text mode issue doesn't replicate. Also, I cannot reproduce an issue with kvm disabled, but then it takes ages to complete installation. With older builds of SLE15 issue was there, but I could resolve it by restarting VM. With this build I can reproduce it in every run. ## steps to reproduce * Download iso using following link for x86-64: https://openqa.suse.de/tests/1149730/asset/iso/SLE-15-Leanos-DVD-x86_64-Build181.4-Media1.iso * Start installation with kvm enabled, e.g.: qemu-system-x86_64 -m 1024 -cpu qemu64 -enable-kvm -hda sle15_181.4.qcow2 \ -cdrom iso/SLE-15-Leanos-DVD-x86_64-Build181.4-Media1.iso * Click Next till Suggested partitioning page * Press expert partitioner button * Try to navigate in the tree on the left side of menu * Qemu freezes, process doesn't respond ## problem Qemu stops responding with SLE 15 iso and kvm enabled. I also was able to reproduce issue with different set of parameters like drive types, cpu, etc.
Small update. Today I've tried upgrading using Virtualization Factory repo, still can reproduce it with qemu-2.10.0-399.7.x86_64. On other box I have it works fine with qemu-2.9.0-2.2.x86_64
Please narrow it down a bit more then. The most common case why VMs stop running is because you're running out of disk space on the host. If this only happens on that one particular machine, it's probably got to do something with that.
Or actually, you're before anything should really write to disk. Can you reproduce this on one machine and give me access to it so I can take a look?
Unfortunately it's my working laptop, so I can only easily provide VNC connection to running VM. I've checked disk space, there is plenty of it available. As I've mentioned above, it works if I select textmode installation of sle-15. Could you please suggest how to narrow down the issue, as I've tried adding verbose logging, but there are no errors in the logs. Strace output: read(32, "g\0\2\0\0\0", 6) = 6 read(32, "\2\0", 2) = 2 read(32, 0x1f99d97ac0, 6) = -1 EAGAIN (Resource temporarily unavailable) ppoll([{fd=3, events=POLLIN}, {fd=5, events=POLLIN}, {fd=7, events=POLLIN}, {fd=8, events=POLLIN}, {fd=13, events=POLLIN}, {fd=16, events=POLLIN}, {fd=24, events=POLLIN}, {fd=25, events=POLLIN}, {fd=26, events=POLLIN}, {fd=27, events=POLLIN}, {fd=28, events=POLLIN}, {fd=29, events=POLLIN}, {fd=32, events=POLLIN}], 13, {tv_sec=0, tv_nsec=622000000}, NULL, 8) = 1 ([{fd=24, revents=POLLIN}], left {tv_sec=0, tv_nsec=173918981}) read(24, "\1\200\302\0\0\0\376T\0\210\226\271\0&BB\3\0\0\0\0\0\200\0RT\0\341\265\377\0\0"..., 69632) = 52 read(24, 0x1f99d257b4, 69632) = -1 EAGAIN (Resource temporarily unavailable) I can also attach debugger to the process, but all I managed to get is [Thread 0x7f524a5ac700 (LWP 14075) exited] [New Thread 0x7f524a5ac700 (LWP 17176)] [Thread 0x7f524a5ac700 (LWP 17176) exited] [New Thread 0x7f524a5ac700 (LWP 24832)] [Thread 0x7f524a5ac700 (LWP 24832) exited] I also have tried different RAM configs, same behavior.
I've tried this scenario using both v2.9 and v2.10 qemu on tumbleweed, and can't reproduce your hang.
I know that there is at least one person who also experiences same issue. So, could you please give a hint how to narrow down possible causes of the issue. I can easily reproduce it. Any ideas?
I can reproduce it and looks like freeze is caused by some kernel driver ( I get this lines using remote logging in installation ) jreidinger@neser-vr:/tmp> cat warn 2017-09-08T12:09:11.084168+00:00 install kernel: [ 412.304271] kernel BUG at ../drivers/gpu/drm/ttm/ttm_bo_util.c:589! 2017-09-08T12:09:11.085165+00:00 install kernel: [ 412.304775] RIP: ttm_bo_kmap+0x1dc/0x230 [ttm] RSP: ffffc90002117c18 I hope it helps you.
After Josef's findings I tried other video devices, and I can reproduce issue only with qxl-vga. With VGA, virtio-vga, VMVGA it works just fine. I hope it helps.
(In reply to Josef Reidinger from comment #7) > I can reproduce it and looks like freeze is caused by some kernel driver ( I > get this lines using remote logging in installation ) > > jreidinger@neser-vr:/tmp> cat warn > 2017-09-08T12:09:11.084168+00:00 install kernel: [ 412.304271] kernel BUG > at ../drivers/gpu/drm/ttm/ttm_bo_util.c:589! > 2017-09-08T12:09:11.085165+00:00 install kernel: [ 412.304775] RIP: > ttm_bo_kmap+0x1dc/0x230 [ttm] RSP: ffffc90002117c18 > > I hope it helps you. This does seem to point to an issue which is already being addressed in boo#1057241. Marking as dup. *** This bug has been marked as a duplicate of bug 1057241 ***