Can you update to Acton, which is released yesterday.
Read More...

upstream libvirt-java still has the issue.
So I grab the upstream libvirt-java 0.4.7, patch it and make a new jar file in git.cloud.com/cgit/cloudstack-oss/tree/d...d-libvirt-0.4.5.jar, which will be included in our next release.
So you can either upgdate your libvirt-java with git.cloud.com/cgit/cloudstack-oss/tree/d...ud-libvirt-0.4.5.jar
or wait for our next release
or downgrade libvirt to 0.8.*
Read More...

It's a bug in libvirt-java, libvir: Storage error : invalid argument: storagePoolRefresh: unsupported flags (0x285a2324)

Basically, in libvirt 0.9.*, they change the storagepoolrefresh api, but libvirt-java is not updated. So here is what happened:
1. mgt server copy one of template from secondary storage to primary storage on one of kvm host, let's say host A.
2. mgt server send a command to kvm host B to create a volume based on the template on that primary storage.
3. host B can't find the template, as it's just created on host A, need to tell libvirt to refresh storage pool.
4. storagepoolrefresh failed, so libvirt can't find that template, then can't create a vm.

It's been fixed in the latest 3.0, which will be released in the these days, please wait and try. Or stay on Centos 6.1.

marigan wrote:

Read More...

Oh, sorry, I haven't get your log yet? Can you send me the log via edison.su At citrix.com? Thanks!
Read More...

Hey, is it possible to set the log level to debug on all kvm hosts, edit /etc/cloud/agent/log4j-cloud.xml, then reproduce the issue from scratch? I need the whole agent/management log?
Read More...

In 2.2.14, we only tested on centos 6.1. So ah, maybe centos 6.2 has the issue with 2.2.14.
Can you try the centos 6.1?
Read More...

the storage sounds good.
1. Can you set the log level to debug in your /etc/cloud/agent/log4j-cloud.xml, replace INFO to DEBUG. Then reproduce the issue.
2. Then fire a bug, and attach the whole mgt server log and agent.log?
Read More...

Seems your libvirtd and cgroup on the destination host have some problems.
Next time, if you met the same issue, please check the cgroup is mounted on the destination host.
Read More...