diff options
Diffstat (limited to '')
| -rw-r--r-- | results/classifier/108/other/1288 | 24 | ||||
| -rw-r--r-- | results/classifier/108/other/1288259 | 67 | ||||
| -rw-r--r-- | results/classifier/108/other/1288620 | 181 |
3 files changed, 272 insertions, 0 deletions
diff --git a/results/classifier/108/other/1288 b/results/classifier/108/other/1288 new file mode 100644 index 000000000..f010774fb --- /dev/null +++ b/results/classifier/108/other/1288 @@ -0,0 +1,24 @@ +graphic: 0.881 +device: 0.821 +other: 0.669 +semantic: 0.560 +vnc: 0.525 +PID: 0.490 +files: 0.478 +performance: 0.405 +boot: 0.383 +debug: 0.374 +socket: 0.335 +permissions: 0.305 +network: 0.276 +KVM: 0.062 + +GPU passing through guest crashes +Description of problem: +First and foremost, I don't know if this is a QEMU, KVM or GPU driver issue. +I began emailing libvirt project and they advised me to contact you, then KVM and then GPU driver developer(NVIDIA). +Host is crashing from time to time. I have guest's kernel dumps(~2GB each). +Steps to reproduce: +Unfortunately, I don't have steps to reproduce. +Additional information: +I'm aware I'm not running the latest qmeu version but I'm willing to install developer version and try to reproduce or test patch if developer requires it. diff --git a/results/classifier/108/other/1288259 b/results/classifier/108/other/1288259 new file mode 100644 index 000000000..312396c72 --- /dev/null +++ b/results/classifier/108/other/1288259 @@ -0,0 +1,67 @@ +KVM: 0.863 +graphic: 0.620 +device: 0.439 +other: 0.418 +performance: 0.348 +socket: 0.341 +vnc: 0.324 +PID: 0.297 +debug: 0.275 +semantic: 0.261 +network: 0.223 +permissions: 0.181 +boot: 0.179 +files: 0.115 + +KVM vms are paused and cannot be deleted due to hardware error 0x0 + +Upon creation of instances via OpenStack nova api instances got stuck in spawning state. Then, after trying to delete instances they got stuck in deleting state. After investigation the following error was found: + +KVM: entry failed, hardware error 0x0 +EAX=00000000 EBX=00000000 ECX=00000000 EDX=00000623 +ESI=00000000 EDI=00000000 EBP=00000000 ESP=00000000 +EIP=0000fff0 EFL=00000002 [-------] CPL=3 II=0 A20=1 SMM=0 HLT=0 +ES =0000 00000000 0000ffff 00009300 +CS =f000 000f0000 0000ffff 0000f300 +SS =0000 00000000 0000ffff 0000f300 +DS =0000 00000000 0000ffff 00009300 +FS =0000 00000000 0000ffff 00009300 +GS =0000 00000000 0000ffff 00009300 +LDT=0000 00000000 0000ffff 00008200 +TR =0000 00000000 0000ffff 00008b00 +GDT= 00000000 0000ffff +IDT= 00000000 0000ffff +CR0=60000010 CR2=00000000 CR3=00000000 CR4=00000000 +DR0=0000000000000000 DR1=0000000000000000 DR2=0000000000000000 DR3=0000000000000000 +DR6=00000000ffff0ff0 DR7=0000000000000400 +EFER=0000000000000000 +Code=28 95 66 ba 01 4a 03 00 66 89 d8 66 5b 66 5e e9 15 79 66 c3 <ea> 5b e0 00 f0 30 36 2f 32 33 2f 39 39 00 fc 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 + +All instances were in paused state: +root@node-7:~# virsh list +setlocale: No such file or directory + Id Name State +---------------------------------------------------- + 4 instance-00000004 paused + 5 instance-00000005 paused + 6 instance-00000006 paused + 7 instance-00000007 paused + 8 instance-00000008 paused + 9 instance-00000009 paused + +The only way to delete VM is to reset it and then resume it. After this, VM is deleted properly. +OpenStack version: Havana on Ubuntu 12.04 +KVM version: QEMU emulator version 1.2.0 (qemu-kvm-1.2.0+noroms-0ubuntu7.12.10, Debian), Copyright (c) 2003-2008 Fabrice Bellard + + + +Status changed to 'Confirmed' because the bug affects multiple users. + +Thanks Thomas for reassigning this to Ubuntu. +That would look like an issue with the KVM HW support, but is too long ago for anything reasonable to be done. + +If this still is an issue please describe what you encounter and add current logs. +Marking incomplete for now. + +[Expired for qemu (Ubuntu) because there has been no activity for 60 days.] + diff --git a/results/classifier/108/other/1288620 b/results/classifier/108/other/1288620 new file mode 100644 index 000000000..be22b74b7 --- /dev/null +++ b/results/classifier/108/other/1288620 @@ -0,0 +1,181 @@ +semantic: 0.642 +other: 0.641 +permissions: 0.577 +debug: 0.546 +graphic: 0.530 +PID: 0.530 +network: 0.527 +vnc: 0.520 +device: 0.446 +KVM: 0.435 +files: 0.432 +performance: 0.430 +socket: 0.353 +boot: 0.332 + +memory leak with config file + +I have a Windows 7 SP1 Professional 64-bit installation on a QCOW2 image with compat=1.1, which I launch via + +qemu-system-x86_64 -drive file=windows_base_HDD.img,index=0,media=disk -enable-kvm -m 512M -vga std -net nic,vlan=0 -net user,vlan=0 + +As soon as I start using the network in any application — for example, visiting www.google.com in Internet Explorer — QEMU starts gobbling memory until the (host) kernel kills it because of an OOM condition. If I run the QEMU with the same options, but with model=e1000 option set for the NIC (i.e. -net -nic,vlan=0,model=e1000), I can use the network from the guest OS without any noticeable effect on QEMU's memory consumption. + +I do not have this problem when running QEMU with the exact same options (as above, without model=e1000) but with a Debian wheezy installation (on a QCOW image of the same format). My host system in Ubuntu 13.10 x86_64, kernel image 3.11.0-17-generic, but with the QEMU packages from trusty (the codename for the next release): +Output of `dpkg -l \*qemu\* | grep '^ii'`: +ii ipxe-qemu 1.0.0+git-20130710.936134e-0ubuntu1 all Virtual package to support use of kvm-ipxe with qemu +ii qemu-keymaps 1.7.0+dfsg-3ubuntu2 all QEMU keyboard maps +ii qemu-system-common 1.7.0+dfsg-3ubuntu2 amd64 QEMU full system emulation binaries (common files) +ii qemu-system-x86 1.7.0+dfsg-3ubuntu2 amd64 QEMU full system emulation binaries (x86) +ii qemu-utils 1.7.0+dfsg-3ubuntu2 amd64 QEMU utilities + +(If necessary, I can try to reproduce this with QEMU built from the upstream source or the latest source from version control.) + +Please try to reproduce this with a debug built (configure --enable-debug) from latest QEMU. If it shows the same memory leak, you can try to find the cause of this leak with valgrind: + +valgrind --leak-check=full qemu-system-x86_64 -drive file=windows_base_HDD.img,index=0,media=disk -enable-kvm -m 512M -vga std -net nic,vlan=0 -net user,vlan=0 2>&1 | tee valgrind.log + +As soon as you terminate the running qemu process, valgrind will write a protocol of all allocated memory. Leaks can usually be found quite easily in this protocol: after a long run, the leak will dominate any other memory allocation. If your executable still contains the full debug information, you will also see the exact line of source code which allocates repeatedly memory. + +Please note that valgrind slows down your process by a factor of 10 to 20, so it takes some time to run Windows. + +If valgrind does not work (which sometimes happens), you can attach a debugger (gdb) to the running qemu process and try to detect the buggy memory allocation by setting breakpoints on the memory allocator functions (malloc or g_malloc, g_malloc0, g_new, g_new0). + + +I can not reproduce this even with the Ubuntu package after rebooting after a kernel update. I will try again with the previous kernel image to confirm this is the relevant variable. Is it possible that the behaviour I described in the initial report is/was caused by code in the KVM module? + +Even after rebooting with the kernel I was using when I had the problem behaviour in QEMU, I can not reproduce the issue. It certainly was not a one-off, because QEMU was gobbling memory consistently on my system, in consecutive sessions. + +I have been able to consistently reproduce the bug again, and have run QEMU with Valgrind until OOM. It is unrelated to networking; it is caused by loading a config file. + +I ran QEMU from Git commit 7f6613cedc59fa849105668ae971dc31004bca1c under valgrind via... + +valgrind qemu-system-x86_64 -readconfig windows8_throwaway_VM.conf -m 1G -vga std 2>&1 | tee valgrind.log + +...where the contents of windows8_throwaway_VM.conf is... + +[drive] + file = "windows8_throwaway_HDD.img" + index = "0" + media = "disk" + if = "virtio" + +[net] + type = "nic" + vlan = "0" + model = "virtio" + +[net] + type = "user" + vlan = "0" + +[rtc] + base = "localtime" + +[machine] + accel = "kvm" + +(I will attach the file in a separate comment, because launchpad appears to only allow at most one attachment per comment.) + +It does not seem to matter whether VirtIO is used, as I have had this problem when not using any VirtIO devices, but the Windows guest I had on-hand was already using it. + +If I invoke QEMU with the equivalent settings all via the command line, it does not gobble memory (again, regardless of VirtIO). + +qemu-system-x86_64 -drive file=windows8_throwaway_HDD.img,index=0,media=disk,if=virtio -enable-kvm -m 1G -vga std -net nic,vlan=0,model=virtio -net user,vlan=0 -rtc localtime + +Attaching config file mentioned in previous comment. + +On Sat, Mar 29, 2014 at 03:02:23AM -0000, Aidan Gauland wrote: +> I have been able to consistently reproduce the bug again, and have run +> QEMU with Valgrind until OOM. It is unrelated to networking; it is +> caused by loading a config file. +> +> I ran QEMU from Git commit 7f6613cedc59fa849105668ae971dc31004bca1c +> under valgrind via... +> +> valgrind qemu-system-x86_64 -readconfig windows8_throwaway_VM.conf -m 1G +> -vga std 2>&1 | tee valgrind.log +> +> ...where the contents of windows8_throwaway_VM.conf is... +> +> [drive] +> file = "windows8_throwaway_HDD.img" +> index = "0" +> media = "disk" +> if = "virtio" +> +> [net] +> type = "nic" +> vlan = "0" +> model = "virtio" +> +> [net] +> type = "user" +> vlan = "0" +> +> [rtc] +> base = "localtime" +> +> [machine] +> accel = "kvm" +> +> (I will attach the file in a separate comment, because launchpad appears +> to only allow at most one attachment per comment.) +> +> It does not seem to matter whether VirtIO is used, as I have had this +> problem when not using any VirtIO devices, but the Windows guest I had +> on-hand was already using it. +> +> If I invoke QEMU with the equivalent settings all via the command line, +> it does not gobble memory (again, regardless of VirtIO). +> +> qemu-system-x86_64 -drive +> file=windows8_throwaway_HDD.img,index=0,media=disk,if=virtio -enable-kvm +> -m 1G -vga std -net nic,vlan=0,model=virtio -net user,vlan=0 -rtc +> localtime + +So this is a problem that only happens under Valgrind? Perhaps this is +a valgrind bug. + +Stefan + + +On Wed, 23 Apr 2014 13:10:39 -0000, Stefan Hajnoczi wrote: +> So this is a problem that only happens under Valgrind? Perhaps this +> is +> a valgrind bug. + +No, it happens outside of Valgrind as well. It only happens when QEMU +is told to read a config file (with -readconfig). + + + +On Wed, Apr 23, 2014 at 08:18:21PM -0000, Aidan Gauland wrote: +> On Wed, 23 Apr 2014 13:10:39 -0000, Stefan Hajnoczi wrote: +> > So this is a problem that only happens under Valgrind? Perhaps this +> > is +> > a valgrind bug. +> +> No, it happens outside of Valgrind as well. It only happens when QEMU +> is told to read a config file (with -readconfig). + +Weird, I tried yesterday and couldn't reproduce it against +qemu.git/master (2d03b49c3f225994c4b0b46146437d8c887d6774) with your +config file. + +I wonder if your guest is repeatedly doing something that causes QEMU to +leak memory. My guest was Red Hat Enterprise Linux 6.4. + +Does it happen if you provide a non-bootable disk image so the guest is +stuck at the BIOS screen? Use dd if=/dev/zero of=test.img bs=1M +count=1024 to create an empty 1 GB raw file. + +Stefan + + +It does seem to be related to the guest, because with a dummy (non-bootable, garbage data) disk image, the rapid memory leak does not occur. + +Can you still reproduce this issue with the latest version of QEMU (currently version 2.9)? + +[Expired for QEMU because there has been no activity for 60 days.] + |