Somehow I got a large number of cinder volumes in available status but have an instance ID in its "attached to". The instance of the instance ID has been deleted in nova. I know this is a bug to enter that state. But with it, how can I delete the volume in this state? By the way, I have tried 'force-delete' and that does not allow me to delete it.
So, there is a workaround for doing this via mysql db, but i would not recommend directly to make changes in DB. But that works.
Step 1: Get volume uuid by running cinder list command
[root@controller ~(openstack_admin)]#cinder list
Step2: You can check the available status and try to reset the state of volume, if that shows "Error_deleting" or "Detaching" Then you can reset-state of the volume from
[root@controller ~(openstack_admin)]#cinder reset-state --state available $volume_uuid
If that also fails, Login to mysql db and use cinder DB
mysql> use cinder;
-following cinder mysql query sets the Cinder state to available:
mysql>update volumes set attach_status='detached',status='available' where id ='$volume_uuid';
or you can also specify more options like setting up instance uuid to NULL and attach time to NULL.
mysql>update volumes set attach_status='detached',instance_uuid=NULL,attach_time=NULL,status="available" WHERE id='volume_uuid';
Step 3: Now again try deleting the effected volume
[root@controller~(openstack_admin)]#cinder delete $volume_uuid
Step 4: if the above workflow dont help then below mysql query will delete the volume
mysql>update volumes set deleted=1,status='deleted',deleted_at=now(),updated_at=now() where deleted=0 and id='$volume_uuid';
In addition to this I;d add following 2 mysql requests:
[cinder]> delete from volume_admin_metadata where volume_id = '$volume_uuid';
[cinder]> delete from volume_attachment where volume_id = '$volume_uuid';
Resetting the volume state won't remove any orphaned attachment records for the volume. It's better to use the force detach API (defaults to admin-only):
https://developer.openstack.org/api-r...
Unfortunately there are no CLIs for this in python-cinderclient or python-openstackclient yet.
Hi,
Can you share the curl procedure? For some reason my cinder refuses to respond properly to curl, and I cannot sort out why ... http://paste.openstack.org/show/729696/
Hi, We don't need to update the mysql database table. We can do as below steps.
Step 1: Make sure that you are admin
# source keystone_admin
Step 2-: Get the UUID of the of the volume you need to remove
# cinder list --all
Step 3-: Reset the state to available
# cinder reset-state --state available <UUID>
Step 4-: detached the volume
# cinder reset-state --attach-status detached <UUID>
Step 5-: Delete the volume now
# cinder delete <UUID>
Hope this is fixed now in 2018, way back in 2015 the steps given by you were not working and thats why we adopted the workaround. -Thanks
I can confirm this does work on a stack running Queens. Thank you for this, was happy to not have to go into the DB.
I've made a script to do so. https://gist.github.com/cygmris/339cc3945072101331b214a8c21b9d9f (delete_volume_attached_with_deleted_instance.sh)
chmod +x delete_volume_attached_with_deleted_instance.sh
./delete_volume_attached_with_deleted_instance.sh ip username password
Asked: 2015-05-15 03:26:44 -0500
Seen: 40,469 times
Last updated: Sep 02 '18
