Ask Your Question

R0peE's profile - activity

2016-04-08 18:02:44 -0500 received badge  Nice Question (source)
2013-11-26 08:23:17 -0500 commented question qemu poor disk performance

I've made some progress on the issue, devstack script packege builds up an openstack instance in my virtualbox vm that has qemu as virtualization, and it uses accel=tcg, i've google around for 3 hours straight and i cant find relevant info on this. What does it do how do i set it up, why is it good?

2013-11-20 04:10:16 -0500 commented answer qemu poor disk performance

Sry for the late response, work was intense and haven't got enough time for debugging. I tried all caching modes, the speed is still the same, around 6MB/sec. Could this be a virtio issue? Oh and I also tried LVM disk instead of disk image, does not effect performance...

2013-11-14 08:47:40 -0500 received badge  Famous Question (source)
2013-11-09 12:48:13 -0500 received badge  Notable Question (source)
2013-11-09 02:23:26 -0500 received badge  Popular Question (source)
2013-11-08 10:01:13 -0500 commented question qemu poor disk performance

I also tried LVM as local disk image source, but it's the same result, as with qcow2 images...

2013-11-08 08:59:12 -0500 asked a question qemu poor disk performance

On one of my compute nodes I had to install qemu as a virtualization option without kvm, bacuse cpu does not support HW virtualization. My problem is, that on the instances the disk performance is terribly slow.

For example, on the host machine doing a 1gig DD looks like this:

time dd if=/dev/zero of=1g bs=4096
count=250000 250000+0 beolvasott
rekord 250000+0 kiírt rekord
1024000000 bájt (1,0 GB) másolva,
2,09917 mp, 488 MB/s

real    0m2.101s user    0m0.040s sys 
0m1.232s

And on the same machine within the instance it looks like this:

time dd if=/dev/zero of=1g bs=4096
count=250000 250000+0 records in
250000+0 records out 1024000000 bytes
(1.0 GB) copied, 165.577 s, 6.2 MB/s

real    2m46.219s user    0m0.688s sys
1m49.931s

this is terrible.. is there any option or tweak to speed things up? my nova-compute.conf looks like this:

[DEFAULT] 
libvirt_type=qemu
libvirt_ovs_bridge=br-int
libvirt_vif_type=ethernet
libvirt_vif_driver=nova.virt.libvirt.vif.LibvirtHybridOVSBridgeDriver
libvirt_use_virtio_for_bridges=True
disk_cachemodes=file=writeback
force_raw_images=True

and the instance start with these parameters:

/usr/bin/qemu-system-x86_64 -name instance-0000006d -S -M pc-1.0 -cpu core2duo,+lahf_lm,+sse4.1,+pdcm,+xtpr,+cx16,+tm2,+est,+ds_cpl,+dtes64,+pbe,+tm,+ht,+ss,+acpi,+ds -no-kvm -m 4096 -smp 2,sockets=2,cores=1,threads=1 -uuid da3956ed-04da-4f35-8b42-fdf50c8e9aa1 -smbios type=1,manufacturer=OpenStack Foundation,product=OpenStack Nova,version=2013.1.3,serial=6825f728-7c97-dd11-88a8-0baf787357d7,uuid=da3956ed-04da-4f35-8b42-fdf50c8e9aa1 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/instance-0000006d.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -no-shutdown -device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 -drive file=/var/lib/nova/instances/da3956ed-04da-4f35-8b42-fdf50c8e9aa1/disk,if=none,id=drive-virtio-disk0,format=qcow2,cache=writeback -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1 -drive file=/var/lib/nova/instances/da3956ed-04da-4f35-8b42-fdf50c8e9aa1/disk.swap,if=none,id=drive-virtio-disk1,format=qcow2,cache=writeback -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x5,drive=drive-virtio-disk1,id=virtio-disk1 -netdev tap,fd=24,id=hostnet0 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=fa:16:3e:f8:0a:07,bus=pci.0,addr=0x3 -chardev file,id=charserial0,path=/var/lib/nova/instances/da3956ed-04da-4f35-8b42-fdf50c8e9aa1/console.log -device isa-serial,chardev=charserial0,id=serial0 -chardev pty,id=charserial1 -device isa-serial,chardev=charserial1,id=serial1 -device usb-tablet,id=input0 -vnc 0.0.0.0:0 -k en-us -vga cirrus -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x6

I would be realy gratefull if someone could help me with this. Thanks in advance!

2013-11-08 08:26:06 -0500 received badge  Famous Question (source)
2013-10-01 14:12:45 -0500 commented question default security group not applied on instance start

ah! I see, yes i followed it, but had to make some adjustments, but I dont realy see where quantum security groups kick in, beacause if i make adjustments on the dashboard, it sends out commands for nova, and quantum is used for tunneling and vlans, but i have to read on this a little bit more

2013-10-01 13:37:04 -0500 commented question default security group not applied on instance start

thats ok for me :) we dont use neutron and the network layout is realy simple

2013-10-01 13:36:19 -0500 answered a question default security group not applied on instance start

thats ok for me :) we dont use neutron and the network layout is realy simple

2013-10-01 13:22:31 -0500 received badge  Notable Question (source)
2013-10-01 13:01:14 -0500 commented question default security group not applied on instance start

It got resolved, i had comment out in /etc/nova.conf: security_group_api=quantum and still there I had to enable: firewall_driver=nova.virt.libvirt.firewall.IptablesFirewallDriver thank you for the help, your tip was a life saver :)

2013-09-26 17:42:28 -0500 received badge  Student (source)
2013-09-26 11:37:20 -0500 received badge  Popular Question (source)
2013-09-26 11:14:39 -0500 commented question default security group not applied on instance start

Yes, I applied it, but after that blackout.., I'm googleing for this problem almost a week now, and no solution, not even a similar case, so I could start debuging.

2013-09-26 10:18:04 -0500 commented question default security group not applied on instance start

Thank you for the comment, but if I do what they suggest, there is absoluteley no connection, with or without the default group, it's like it don't even exists.. :-(

2013-09-25 18:12:31 -0500 received badge  Editor (source)
2013-09-25 18:00:00 -0500 asked a question default security group not applied on instance start

My main problem is that if I launch an instance, the default security group is not applied even though it is checkd in by default. The other problem is, when the instance is launched, on my compute nodes i loose connectivity to the instances. I have to swap out then swap in the security group, and everything is okay. The instances start normally, they get ip address, they have connectivity.

I used this guide to set up my openstack multinode installation: https://github.com/mseknibilel/OpenStack-Grizzly-Install-Guide/blob/OVS_MultiNode/OpenStack_Grizzly_Install_Guide.rst

This is my quantum.conf: http://pastebin.com/YhwBpNca

And this is my ovs_quantum_plugin.ini: http://pastebin.com/N0bJ2z4B

And for the end nova.conf network stuff: http://pastebin.com/t92rjvk4