While running "cf copy-source APP1 APP2", users see the following error:
FAILED Failed to perform blobstore operation after three retries.
From cloud controller server /var/vcap/sys/log/cloud_controller_ng/cloud_controller_ng.log
{"timestamp":1488480555.86369,"message":"cp_file_between_keys-retry","log_level":"debug","source":"cc.blobstore.dav_client","data":{"request_guid":"c1674444 -cc22-4acd-572b-01b6467b1b22::1232a487-9a34-4565-a65d-ccb87da14477","error":"SSL verification failed","remaining_retries":0,"args":{"source_key":"65d57057-d 1e6-4766-a97c-89e976180770","destination_key":"b966ad4b-18a1-4072-a3ba-fe99ddd7b09f"}},"thread_id":46914699153700,"fiber_id":46914742844680,"process_id":118 45,"file":"/var/vcap/data/packages/cloud_controller_ng/5f59122e1ae729a0ec63a4d511cf61f284df9e97.1-6e01d314dfb3752e6b32c8a467467889309c34b7/cloud_controller_ ng/lib/cloud_controller/blobstore/retryable_client.rb","lineno":131,"method":"rescue in with_retries"}
This is a known issue that is causing the failure of the background jobs that interact with WebDAV. The issue also results in blobs not being deleted. The fix is being worked on now and will be backported to 1.8, 1.9. Pivotal does not currently have a procedure available for reclaiming the lost space. If this error has caused you to lose a large amount of space, please contact Pivotal Support.
cloud_controller_worker config:
The file /var/vcap/jobs/cloud_controller_worker/config/cloud_controller_ng.yml
referenced the following for certificates for the blobstores:
ca_cert_path: /var/vcap/jobs/cloud_controller_ng/config/certs/buildpacks_ca_cert.pem
But this is the path for cloud controller, not the cloud controller worker. The correct path for the worker should be:
ca_cert_path: /var/vcap/jobs/cloud_controller_worker/config/certs/buildpacks_ca_cert.pem
This is causing some worker tasks to fail with cert errors when accessing blobstore endpoints.
The fix is available in PCF version 1.8.33 and version 1.9.6 onwards.