Cinder volume starting for very long time

[Michal Arbet]

Hi,

We have victoria deployment where when we restart cinder-volume with rbd backends it takes 25 minutes to start cinder-volume. This is caused by cinder-volume which is asking ceph via rbd "how big is this volume". In our case we have 50 000 volumes in ceph.

If rbd backend (pool) in ceph is used only for openstack, there is no reason to ask ceph for size for each volume and better is to let cinder trust his DB.

Because of above cinder proposed configration option below :

# Set to False if the pool is shared with other usages. On exclusive use driver
# won't query images' provisioned size as they will match the value calculated
# by the Cinder core code for allocated_capacity_gb. This reduces the load on
# the Ceph cluster as well as on the volume service. On non exclusive use
# driver will query the Ceph cluster for per image used disk, this is an
# intensive operation having an independent request for each image. (boolean
# value)
#rbd_exclusive_cinder_pool = true

This patch is merged in master and backported until rocky (and higher).
https://review.opendev.org/q/I32c7746fa9149bce6cdec96ee9aa87b303de4271

We also found that we are using older stable/victoria for cinder image (default rbd_exclusive_cinder_pool from code is false in our version) - so solution is also to upgrade cinder images.

But this should be definitively configurable via kolla-ansible as other rbd_* options, moreover if this value was changed in mid cycle.

[Jon Bernard (jbernard)]

This configuration option was made the default and backported to the Victoria release in May of last year.

https://review.opendev.org/q/I839441a71238cdad540ba8d9d4d18b1f0fa3ee9d

I'm not sure there's much that Cinder can do here, maybe an update to the latest version, or filing a bug with kolla-ansible.

[Michal Arbet (michalarbet)]

Jon, thanks for reply, I wanted to create bug in kolla-ansible project not in cinder, sorry.