Going back to the original question about how much does this improve performance, here's one result. I had two containers in AWS which had ~84400 blobs in each. I deleted the first container with the existing jclouds code and the second container with the new code. See results below.
``` $ ./bin/jclouds blobstore list shri-nov6-3-cloud-testing0-1 | wc -l 84389 $ time ./bin/jclouds blobstore container-delete shri-nov6-3-cloud-testing0-1 real 24m59.975s user 4m44.908s sys 1m21.324s $ ./bin/jclouds blobstore list shri-nov6-3-cloud-testing0-2 | wc -l 84387 $ time ./bin/jclouds blobstore container-delete shri-nov6-3-cloud-testing0-2 real 10m52.382s user 4m39.216s sys 1m17.214s ``` I built jclouds-cli for these tests. To make sure jclouds-cli was using my jclouds jar, this is the way I built it (as per the suggestion on the irc channel). 1. run mvn install in jclouds 2. run mvn install in jclouds-karaf 3. run mvn package in jclouds-cli which created the tar.gz that was extracted and run. --- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds/pull/214#issuecomment-29853486
