On 03/27/2014 03:52 AM, Wanghao (S) wrote: > Hi, all, > > > > There is a use case: we have two nova components (call them nova A and > nova B) and one cinder component. Attach a volume to an instance in nova > A and then services of nova A become abnormal. > > Because the volume also want to be used in nova B, so using cinder api > "force detach volume" to free this volume. But when nova A is normal, > nova can't detach this volume from instance by using nova api "detach > volume" , > > as nova check the volume state must be "attached". > > > > I think should we add "force detach" function to nova just like "attach" > and "detach", because if using force detach volume in cinder, there is > still some attach information in nova which can't be cleaned by using > nova api "detach". > > > > There is the BP link > ：https://blueprints.launchpad.net/nova/+spec/add-force-detach-to-nova >
Hi, Please be aware that we are changing how BPs are done in Juno. You can see more details in this email thread [1]. Also as mentioned on the bug that started this [2], the reason I think this needs a BP is because there are edge cases to be discussed and figured out - not only because we need to follow a process. Addressing some of the concerns from the bug on the gerrit proposal would be great. Thanks, N. [1] http://lists.openstack.org/pipermail/openstack-dev/2014-March/030576.html [2] https://bugs.launchpad.net/nova/+bug/1297127 > > > Any suggestion is great. THX~ > > > > Sorry for the first email without subject, please ignore it. > > > > _______________________________________________ > OpenStack-dev mailing list > OpenStack-dev@lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > _______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
