object save code is terribly inefficient

[Alex Schultz]

I attempted to download a 10M file using openstack object save <container> <file> from a swift instance on localhost and it sat chunking the response in the following loop in the following bit of code for some time:

https://github.com/openstack/python-openstackclient/blob/b55b1d2b94c7ef668ab06b4047a7bf2209d9b275/openstackclient/api/object_store_v1.py#L382-L384

It might be better to just let the underlying curl write out the file somewhere since when I debugged it, the curl was done almost immediately but the writing out took in the order of minutes to write out the file with openstack client pegged at 100% cpu.

[Alex Schultz (alex-schultz)]

It should be noted it took 4 minutes to process the 10M file.

time openstack object save --file test.tar.xz overcloud_support sosreport-overcloud-controller-0-20170113221415.tar.xz

real 4m1.562s
user 3m58.055s
sys 0m1.730s

[Steve Martinelli (stevemar)]

This was solved in https://review.openstack.org/#/c/417892/
and is a dupe of https://bugs.launchpad.net/python-openstackclient/+bug/1654645

Please note that it depends on the latest keystoneauth to work