"lyz_pro"<lyz_pro@163.com>
2014-Apr-11 08:02 UTC
[libvirt-users] libvirt with glusterfs problem
Hi
I have two node: hcg3, hcg4 which run glusterfs
Also run libvirtd on each node.
The problem cames when I use `virsh attach-device` to attach a glusterfs type
disk .
My device xml like:
<disk type="network" device="disk">
<driver name="qemu" type="raw"
cache="none"/>
<source protocol="gluster"
name="h2312v3/volume-257e07bd-9e80-4d06-8d6d-5f83979de1fd">
<host name="hcg4"/>
</source>
<target bus="virtio" dev="vdb"/>
<serial>257e07bd-9e80-4d06-8d6d-5f83979de1fd</serial>
</disk>
when I attach a gluster disk to a domain running on hcg4.
the above xml work great.
BUT when I try to change the tag `host` name attribute to hcg3, the attach
action will fail with output:
error: Requested operation is not valid: domain is not running
Also:
I can use `qemu-kvm` command to start vm successful with gluster://hcg3 or
gluster://hcg4.
I have stop both nodes' iptables. and my libvirt and qemu version is :
Compiled against library: libvirt 1.1.4
Using library: libvirt 1.1.4
Using API: QEMU 1.1.4
Running hypervisor: QEMU 1.6.2
glusterfs-3.4.1-3.el6.x86_64
--------------
lyz_pro
2014-04-11
"lyz_pro"<lyz_pro@163.com>
2014-Apr-11 08:10 UTC
Re: [libvirt-users] libvirt with glusterfs problem
Sorry, My mistake. the libvirt error output is:
error: Timed out during operation: cannot acquire state change lock
------------------
lyz_pro
2014-04-11
-------------------------------------------------------------
发件人："lyz_pro"<lyz_pro@163.com>
发送日期：2014-04-11 16:02
收件人：libvirt-users
抄送：
主题：[libvirt-users] libvirt with glusterfs problem
Hi
I have two node: hcg3, hcg4 which run glusterfs
Also run libvirtd on each node.
The problem cames when I use `virsh attach-device` to attach a glusterfs type
disk .
My device xml like:
<disk type="network" device="disk">
<driver name="qemu" type="raw"
cache="none"/>
<source protocol="gluster"
name="h2312v3/volume-257e07bd-9e80-4d06-8d6d-5f83979de1fd">
<host name="hcg4"/>
</source>
<target bus="virtio" dev="vdb"/>
<serial>257e07bd-9e80-4d06-8d6d-5f83979de1fd</serial>
</disk>
when I attach a gluster disk to a domain running on hcg4.
the above xml work great.
BUT when I try to change the tag `host` name attribute to hcg3, the attach
action will fail with output:
error: Requested operation is not valid: domain is not running
Also:
I can use `qemu-kvm` command to start vm successful with gluster://hcg3 or
gluster://hcg4.
I have stop both nodes' iptables. and my libvirt and qemu version is :
Compiled against library: libvirt 1.1.4
Using library: libvirt 1.1.4
Using API: QEMU 1.1.4
Running hypervisor: QEMU 1.6.2
glusterfs-3.4.1-3.el6.x86_64
--------------
lyz_pro
2014-04-11
_______________________________________________
libvirt-users mailing list
libvirt-users@redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-users
On 04/11/2014 02:10 AM, lyz_pro wrote:> Sorry, My mistake. the libvirt error output is: > error: Timed out during operation: cannot acquire state change lockI wonder if this is the same bug as what has been raised on the qemu list: https://lists.gnu.org/archive/html/qemu-devel/2014-04/msg02699.html> Running hypervisor: QEMU 1.6.2 > glusterfs-3.4.1-3.el6.x86_64If I'm correct, the bug is happening because the version of qemu you are using calls glfs_fini in the wrong place, where older glusterfs treated the call as a no-op but the newer glusterfs you are using hangs when called in the wrong order. If so, the fix is in qemu, not libvirt. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org