summary refs log tree commit diff stats
path: root/results/classifier/118/hypervisor/1201447
blob: 0bc6c87d3af2d468f8a4ea0d717633d44ab6dfae (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
hypervisor: 0.982
KVM: 0.978
virtual: 0.940
peripherals: 0.924
device: 0.922
files: 0.915
performance: 0.726
architecture: 0.700
PID: 0.685
graphic: 0.669
ppc: 0.663
kernel: 0.651
register: 0.645
semantic: 0.630
mistranslation: 0.623
user-level: 0.618
permissions: 0.537
socket: 0.536
network: 0.462
vnc: 0.459
boot: 0.446
VMM: 0.422
x86: 0.329
debug: 0.312
risc-v: 0.299
TCG: 0.274
assembly: 0.244
arm: 0.220
i386: 0.220

Blue screen when disk uses cache='writeback'

I am running Windows 2008R2 as KVM guest on Ubuntu 12.04 hypervisor. Disk controller and network card are virtio devices with drivers from https://launchpad.net/kvm-guest-drivers-windows/+download (virtio-win-drivers-20120712-1.iso).
Everything worked fine until I changed disk controller cache from the default (writethrough) to writeback. This introduced occasional blue screens. I noticed that they are linked to high disk IO. For example restoring over 1GB of backup files will results in a blue screen on around 4 out of 5 attempts. Also Windows update crashes the system sometimes. When idle the system will run fine for hours or sometimes even days.
After removing cache='writeback' from the config everything came back to normal.

qemu-kvm:
  Installed: 1.0+noroms-0ubuntu14.8
  Candidate: 1.0+noroms-0ubuntu14.8
  Version table:
 *** 1.0+noroms-0ubuntu14.8 0
        500 http://archive.ubuntu.com/ubuntu/ precise-updates/main amd64 Packages
        100 /var/lib/dpkg/status
     1.0+noroms-0ubuntu14.7 0
        500 http://security.ubuntu.com/ubuntu/ precise-security/main amd64 Packages
     1.0+noroms-0ubuntu13 0
        500 http://archive.ubuntu.com/ubuntu/ precise/main amd64 Packages



Hi Jacek,
I haven't seen other writeback related crashes and the info so far isn't enough to debug anything.

Did you get any related host logs on the crashes.
In dmesg or the guest log in /var/log/libvirt/qemu/<guestname>.log?

[Expired for qemu (Ubuntu) because there has been no activity for 60 days.]