Comment 0 for bug 2020139

When Cinder volume service creates a volume from an image, it will download an image from Glance to a cinder-volume unit (to be precise, when CoW is not available by the backend between Glance and Cinder).

The image fetch speed is not super fast as of today:

2023-05-17 02:52:28.275 158391 DEBUG cinder.image.image_utils [req-e4234307-ae7c-4220-a10a-15760679ac01 a32b8e6ed20f484eba8ddae8f7b6eb1e 08dd29cd8676470c8e0b73d2667f8cf0 - - -] Image fetch details: dest /var/lib/cinder/conversion/image_fetch_741953c3-08f7-43b6-b423-2be938ac8d02_bc1hvgd7cinder@cinder-ceph, size 1907.35 MB, duration 27.80 sec fetch /usr/lib/python3/dist-packages/cinder/image/image_utils.py:637
2023-05-17 02:52:28.275 158391 INFO cinder.image.image_utils [req-e4234307-ae7c-4220-a10a-15760679ac01 a32b8e6ed20f484eba8ddae8f7b6eb1e 08dd29cd8676470c8e0b73d2667f8cf0 - - -] Image download 1907.35 MB at 68.61 MB/s

-> ~ 549 Mbps

By running a profiler, it looks like serious_integrity_iter (glanceclient/common/utils.py) is consuming a lot of CPU cycles on the client side (in this case the cinder-volume unit).