[virt-tools-list] C'ant migrate with virt-manager or virsh

[Bruno Lamps]

Hi, /var/log/libvirt/qemu/lol.log didn't look very informative about the
migration error, so I edited /etc/libvirt/libvirtd.conf and add the log
lines.

Before trying to migrate the VM, I remembered that the /etc/hosts was
configured, but iptables was probably running, because I rebooted these
machines like 10 times since yesterday. So I stoped both iptables services,
tryed to migrate the VM using virt-manager and (OMG what a
*beaultiful*moment!!!) and it migrated in like 1
second!!!!!11!1!eleven!!!!! (no OS
running, just the VM powered on, probably 0MB of RAM being used).

I started iptables in source machine, and migration worked again. When I
started iptables in destination computer, migration failed like I was used
to see. Opening libvirtd.log and looking for "migrat", i found this:

Ocurrences to "migrat" in libvirtd.log:

11:23:22.912: error : qemuDomainMigrateSetMaxDowntime:11781 : invalid
> argument in qemuDomainMigrateSetMaxDowntime: unsupported flags (0xbc614e)
> 11:23:22.917: error : qemuDomainMigrateSetMaxDowntime:11781 : invalid
> argument in qemuDomainMigrateSetMaxDowntime: unsupported flags (0xbc614e)
> 11:23:23.049: debug : qemuMonitorGetBalloonInfo:1001 : mon=0x7f266c000ef0
> 11:23:23.049: debug : qemuMonitorCommandWithHandler:231 : Send command
> 'info balloon' for write with FD -1
> 11:23:23.050: debug : qemuMonitorCommandWithHandler:236 : Receive command
> reply ret=0 errno=0 21 bytes 'balloon: actual=300
>
> 11:23:24.513: debug : qemuMonitorMigrateToHost:1312 : mon=0x7f266c000ef0
> hostname=cluster.pisolar port=49160 flags=1
> 11:23:24.513: debug : qemuMonitorCommandWithHandler:231 : Send command
> 'migrate  -d"tcp:cluster.pisolar:49160"' for write with FD -1
> 11:23:24.514: debug : qemuMonitorCommandWithHandler:236 : Receive command
> reply ret=0 errno=0 0 bytes '(null)'
> 11:23:24.514: debug : qemuMonitorGetMigrationStatus:1283 :
> mon=0x7f266c000ef0
> 11:23:24.514: debug : qemuMonitorCommandWithHandler:231 : Send command
> 'info migrate' for write with FD -1
> 11:23:24.515: debug : qemuMonitorCommandWithHandler:236 : Receive command
> reply ret=0 errno=0 26 bytes 'Migration status: failed
>
> 11:23:24.515: error : qemuDomainWaitForMigrationComplete:5119 : operation
> failed: Migration unexpectedly failed
> 11:23:25.028: debug : qemuMonitorGetBalloonInfo:1001 : mon=0x7f266c000ef0
> 11:23:25.028: debug : qemuMonitorCommandWithHandler:231 : Send command
> 'info balloon' for write with FD -1
> 11:23:25.028: debug : qemuMonitorCommandWithHandler:236 : Receive command
> reply ret=0 errno=0 21 bytes 'balloon: actual=300
>


qemu/lol.log, the log file for the machine, only has lines with the VM
configuration wrote, I think a new line like this is written into this file
everytime I run the VM. An example line:

LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin QEMU_AUDIO_DRV=none
> /usr/bin/qemu-kvm -S -M pc-0.13 -enable-kvm -m 300 -smp
> 1,sockets=1,cores=1,threads=1 -name lol -uuid
> eb66789f-20d6-d1c2-f9dc-6a33c62e10a8 -nodefconfig -nodefaults -chardev
> socket,id=monitor,path=/var/lib/libvirt/qemu/lol.monitor,server,nowait -mon
> chardev=monitor,mode=readline -rtc base=utc -boot c -drive
> file=/media/nfs067/lol.img,if=none,id=drive-ide0-0-0,boot=on,format=raw
> -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -drive
> if=none,media=cdrom,id=drive-ide0-1-0,readonly=on,format=raw -device
> ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 -device
> rtl8139,vlan=0,id=net0,mac=52:54:00:5e:36:b0,bus=pci.0,addr=0x3 -net
> tap,fd=37,vlan=0,name=hostnet0 -chardev pty,id=serial0 -device
> isa-serial,chardev=serial0 -usb -vnc 127.0.0.1:0 -vga cirrus -device
> AC97,id=sound0,bus=pci.0,addr=0x4 -incoming exec:cat -device
> virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5
> char device redirected to /dev/pts/1
>

I opened tcp port 49160 in destination machine, tried to migrate the VM, and
the log pointed a problem with port 49161. So I opened it and tryed again,
and log said I had some problem with 49162. So I opened 49160:49168 tcp
ports, and migration worked with both iptables services running. Cool!

Does anyone knows the exact number of necessary ports?


I'm going to try to migrate real VMs now, form different folders, and with
virsh too. 'Will try virsh migrate --live --p2p --tunnelled lol
qemu+ssh://root@cluster.pisolar/system too, of course.

Thank you all guys very much, and again sorry if my english isn't easy and
clear to understand. =D





On Thu, Feb 3, 2011 at 9:25 AM, Justin Clift <jclift at redhat.com> wrote:

> On 03/02/2011, at 10:48 PM, Bruno Lamps wrote:
> <snip>
> > Trying to use virsh, I get this answer:
> >
> > [root at fedorao ~]# virsh migrate --live lol qemu+ssh://192.168.1.4/system
> > root at 192.168.1.4's password:
> > erro:operation failed: Migration unexpectedly failed
>
> Hmmm, just in the category of "see what happens", give this a go and see
> if it makes any difference at all (prob not):
>
>  $ virsh migrate --live --p2p --tunnelled lol
> qemu+ssh://root@cluster.pisolar/system
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/virt-tools-list/attachments/20110203/1f2e2284/attachment.htm>