VDSM sync meeting minutes August 5th, 2014
Yesterdays we discussed various pending issues, and among them: - live merge for 3.5 has backported patches pending merge http://gerrit.ovirt.org/#/q/status:open+project:vdsm+branch:ovirt-3.5+topic:... - jsonrpc inroduced logging Engine command arguments, equivalent to what we have in xmlrpc - We have a ppc-specific bug - on vdsm restart, previously-running VMs are not recovered. This is very dangerous, as it may lead to Engine restarting the same VM elsewhere, causing data curruption. Michal found the bug and its cause - in ppc, we do not inject the vm uuid to the guest bios, and clientIF.isVDSMVm() fails to recognize such Vms. Is there a BZ open on that? - There was an agreement that we configure libvirtd.log to be too verbose. Yaniv/Mooli, please consider something like my untested http://gerrit.ovirt.org/31135 Dan.
On Wed, Aug 06, 2014 at 11:28:14AM +0100, Dan Kenigsberg wrote:
Yesterdays we discussed various pending issues, and among them:
- live merge for 3.5 has backported patches pending merge http://gerrit.ovirt.org/#/q/status:open+project:vdsm+branch:ovirt-3.5+topic:...
- jsonrpc inroduced logging Engine command arguments, equivalent to what we have in xmlrpc
- We have a ppc-specific bug - on vdsm restart, previously-running VMs are not recovered. This is very dangerous, as it may lead to Engine restarting the same VM elsewhere, causing data curruption. Michal found the bug and its cause - in ppc, we do not inject the vm uuid to the guest bios, and clientIF.isVDSMVm() fails to recognize such Vms.
Is there a BZ open on that?
There is: 1126887 recovery of VMs after VDSM restart doesn't work on PPC
- There was an agreement that we configure libvirtd.log to be too verbose. Yaniv/Mooli, please consider something like my untested http://gerrit.ovirt.org/31135
Dan.
participants (1)
-
Dan Kenigsberg