Cinder default configuration, regardless of whether the backend is LVM or ceph delete volume, it will clear 0 of the entire partition, resulting in slow speed.
The log shows that he wrote the disk directly with the DD command.
DEBUG Oslo_concurrency.processutils [req-47e7d88d-8bf1-4cb3-980b-26f713579e8e 2b6c47d8328b4270ab643b3ab9870d54 1cc33477a6594fd2a0228ec16c59f4eb-default default] CMD "sudo cinder-rootwrap/etc/cinder/rootwrap.conf dd count=0 if=/ Dev/zero of=/dev/mapper/cinder--volumes-volume--d271a71f--608d--4291--ab94--30cd4c8cea70 Oflag=direct "returned:0 In 0.127s execute/usr/lib/python2.7/site-packages/oslo_concurrency/processutils.py:374
This means that the deletion time is proportional to the volume size,
If this security is not the case. Consider turning off this feature
Add in config file
Volume_clear = None
The default value is zero
Check the log has not cleared 0 this action. Seconds to delete
2016-11-10 21:45:35.661 175421 INFO CINDER.VOLUME.DRIVERS.LVM [req-d002a265-14b2-4267-9444-3283f5908337 2B6C47D8328B4270AB643B3AB9870D54 1cc33477a6594fd2a0228ec16c59f4eb-default Default] successfully deleted volume: 74e0d582-6fc0-452a-a972-0e02456add052016-11-10 21:45:36.198 175421 INFO Cinder.volume.manager [ req-d002a265-14b2-4267-9444-3283f5908337 2b6c47d8328b4270ab643b3ab9870d54 1cc33477a6594fd2a0228ec16c59f4eb- Default default] Deleted volume successfully.
But remember, the production environment does not recommend this,
Accelerate Cinder Delete volume speed