[OmniOS-discuss] Ang: Re: Ang: Re: Ang: Re: r151014 KVM crash

Johan Kragsterman johan.kragsterman at capvert.se
Mon Apr 6 17:42:23 UTC 2015 

Thnks, Jorge!

Yeah, I'm the one with the problems...

So you're using kvmadm...? Is that from smartOS, huh? How did you manage to get it to work on OmniOS?

Rgrds Johan


-----Jorge Schrauwen <sjorge+ml at blackdot.be> skrev: -----
Till: Johan Kragsterman <johan.kragsterman at capvert.se>
Från: Jorge Schrauwen <sjorge+ml at blackdot.be>
Datum: 2015-04-06 19:25
Kopia: Michael Rasmussen <mir at miras.org>, omnios-discuss at lists.omniti.com
Ärende: Re: Ang: Re: [OmniOS-discuss] Ang: Re: r151014 KVM crash

Who ever is having the problem :)
It was a bit hard to follow, I'm using kvmadm and this is my config:
{
    "cosmos" : {
       "nics" : [
          {
             "index" : "0",
             "nic_name" : "vcosmos0",
             "model" : "virtio",
             "vlan_id" : "10"
          },
          {
             "index" : "1",
             "nic_name" : "vcosmos1",
             "model" : "virtio"
          }
       ],
       "cpu_type" : "qemu64,+aes,+sse4.2,+sse4.1,+ssse3",
       "hpet" : "true",
       "usb_tablet" : "true",
       "disks" : [
          {
             "index" : "0",
             "boot" : "true",
             "disk_path" : "main/vms/hosts/cosmos/disk0",
             "model" : "virtio"
          }
       ],
       "shutdown" : "acpi_kill",
       "boot_order" : "cd",
       "vcpus" : "sockets=1,cores=4,threads=2",
       "serials" : [
          {
             "index" : "0",
             "serial_name" : "console"
          }
       ],
       "cleanup" : "true",
       "ram" : "6144",
       "time_base" : "utc"
    }
}


Which turns into:
-(~)-[!]-{ pfexec pargs 616                                    
}-(sjorge at core)-
616:    /usr/bin/qemu-system-x86_64 -name cosmos -enable-kvm -m 6144 
-cpu qemu64,+aes,+
argv[0]: /usr/bin/qemu-system-x86_64
argv[1]: -name
argv[2]: cosmos
argv[3]: -enable-kvm
argv[4]: -m
argv[5]: 6144
argv[6]: -cpu
argv[7]: qemu64,+aes,+sse4.2,+sse4.1,+ssse3
argv[8]: -smp
argv[9]: sockets=1,cores=4,threads=2
argv[10]: -rtc
argv[11]: base=utc,driftfix=slew
argv[12]: -pidfile
argv[13]: /var/run/kvm/cosmos.pid
argv[14]: -monitor
argv[15]: unix:/var/run/kvm/cosmos.monitor,server,nowait,nodelay
argv[16]: -vga
argv[17]: none
argv[18]: -nographic
argv[19]: -drive
argv[20]: 
file=/dev/zvol/rdsk/main/vms/hosts/cosmos/disk0,if=virtio,media=disk,index=0,cache=none,boot=on
argv[21]: -boot
argv[22]: order=cd
argv[23]: -device
argv[24]: 
virtio-net-pci,mac=02:08:20:f5:95:f4,tx=timer,x-txtimer=200000,x-txburst=128,vlan=10
argv[25]: -net
argv[26]: vnic,vlan=10,name=net0,ifname=vcosmos0
argv[27]: -device
argv[28]: 
virtio-net-pci,mac=02:08:20:0c:08:d2,tx=timer,x-txtimer=200000,x-txburst=128,vlan=0
argv[29]: -net
argv[30]: vnic,vlan=0,name=net1,ifname=vcosmos1
argv[31]: -chardev
argv[32]: 
socket,id=serial0,path=/var/run/kvm/cosmos.console,server,nowait
argv[33]: -serial
argv[34]: chardev:serial0
argv[35]: -usb
argv[36]: -usbdevice
argv[37]: tablet
argv[38]: -daemonize

PS: shout out to dan for introducing me to pgrep and pargs!

Regards

Jorge

On 2015-04-06 19:22, Johan Kragsterman wrote:
> Jorge, I thought you would jump in on this thread!
> 
> Are you asking me or Michael?
> 
> I'm not running ivy bridge on this machine, it's an older westmere...
> 
> Yeah, I pass --cpu=host
> 
> Can you provide your config file, pls?
> 
> Mine is like this:
> 
> 
> 
> 
> root at omni2:/usr/bin# cat vmedu14041.sh
> #!/usr/bin/bash
> 
> # on omnios command is /usr/bin/qemu-system-x86_64
> 
> # configuration
> NAME="EDU14041"
> NUM=4
> VNIC0=ltsp0
> VNIC1=ltsp1
> VNIC2=ltsptl0
> HDD0=/dev/zvol/rdsk/mainpool/vm/edu/os
> HDD1=/dev/zvol/rdsk/mainpool/vm/edu/opt
> CD=/mainpool/iso/edu14041.iso
> HDD2=/dev/zvol/rdsk/mainpool/vm/edu/home
> HDD3=/dev/zvol/rdsk/mainpool/vm/edu/swap
> MEM=8192
> 
> # don't change below here!
> 
> TLN=`expr 7000 + $NUM`
> mac0=`dladm show-vnic -po macaddress $VNIC0`
> mac1=`dladm show-vnic -po macaddress $VNIC1`
> mac2=`dladm show-vnic -po macaddress $VNIC2`
> 
> /usr/bin/qemu-system-x86_64 \
> -name $NAME \
> -boot cd \
> -enable-kvm \
> -vnc 0.0.0.0:$NUM \
> -smp cores=2,threads=2 \
> -m $MEM \
> -no-hpet \
> -localtime \
> -drive file=$HDD0,if=virtio,index=0 \
> -drive file=$HDD1,if=virtio,index=2 \
> -drive file=$CD,media=cdrom,if=ide,index=1 \
> -drive file=$HDD2,if=virtio,index=3 \
> -drive file=$HDD3,if=virtio,index=4 \
> -net nic,vlan=0,name=net0,model=virtio,macaddr=$mac0 \
> -net vnic,vlan=0,name=net0,ifname=$VNIC0,macaddr=$mac0 \
> -net nic,vlan=1,name=net1,model=virtio,macaddr=$mac1 \
> -net vnic,vlan=1,name=net1,ifname=$VNIC1,macaddr=$mac1 \
> -net nic,vlan=2,name=net2,model=virtio,macaddr=$mac2 \
> -net vnic,vlan=2,name=net2,ifname=$VNIC2,macaddr=$mac2 \
> 
> -vga std \
> -cpu host \
> -pidfile /mainpool/vm/edu/pids/$NAME.pid \
> -monitor telnet:localhost:$TLN,server,nowait,nodelay \
> -daemonize
> 
> if [ $? -gt 0 ]; then
>     echo "Failed to start VM"
>     exit
> fi
> 
> port=`expr 5900 + $NUM`
> public_nic=$(dladm show-vnic|grep vnic0|awk '{print $2}')
> public_ip=$(ifconfig $public_nic|grep inet|awk '{print $2}')
> 
> echo "Started VM: $NAME"
> echo "VNC available at: host IP ${public_ip} port ${port}"
> echo "QEMU Monitor, do: # telnet localhost $TLN. Note: use Control ]
> to exit monitor before quit!
> "
> 
> 
> Rgrds Johan
> 
> 
> -----"OmniOS-discuss" <omnios-discuss-bounces at lists.omniti.com> skrev: 
> -----
> Till: Michael Rasmussen <mir at miras.org>
> Från: Jorge Schrauwen
> Sänt av: "OmniOS-discuss"
> Datum: 2015-04-06 19:06
> Kopia: omnios-discuss at lists.omniti.com
> Ärende: Re: [OmniOS-discuss] Ang: Re: r151014 KVM crash
> 
> Out of curiosity, are you passing --cpu=host? I had issues with that on
> my ivy bridge.
> I currently use: qemu64,+aes,+sse4.2,+sse4.1,+ssse3 which seems to make
> a lot of things smoother.
> 
> I stil get these but qemu does not crash for me:
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff21ff94c000, id=0, base_msr= fee00100 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff21ff944000, id=1, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff21ff93c000, id=2, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff2201f14000, id=3, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff2201f0c000, id=4, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff2201f04000, id=5, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff2201efc000, id=6, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> Apr  1 19:47:28 core.acheron.be kvm: [ID 420667 kern.info]
> kvm_lapic_reset: vcpu=ffffff2201f54000, id=7, base_msr= fee00000 PRIx64
> base_address=fee00000
> Apr  1 19:47:28 core.acheron.be kvm: [ID 710719 kern.info] vmcs
> revision_id = 10
> 
> 
> Regards
> 
> Jorge
> 
> 
> On 2015-04-06 18:03, Michael Rasmussen wrote:
>> On Mon, 6 Apr 2015 11:55:27 -0400
>> Dan McDonald <danmcd at omniti.com> wrote:
>> 
>>> I'm talking with the illumos KVM folks.  They mentioned that Ivy
>>> Bridge Xeons (i.e. Xeon E5-26xx v2, where v2 means Ivy Bridge) have
>>> erratum that can cause problems.  That you do not seem to see these 
>>> in
>>> r151012, however, is very odd.
>>> 
>> There is also E3-12xx v2 which is Ivy Bridge based. Are these affected
>> by this erratum as well?
>> 
>> _______________________________________________
>> OmniOS-discuss mailing list
>> OmniOS-discuss at lists.omniti.com
>> http://lists.omniti.com/mailman/listinfo/omnios-discuss
> _______________________________________________
> OmniOS-discuss mailing list
> OmniOS-discuss at lists.omniti.com
> http://lists.omniti.com/mailman/listinfo/omnios-discuss

More information about the OmniOS-discuss mailing list