Hide Forgot
Description of problem: The ceph rbd checker uses hundreds of MB per image/device. This causes the OOM to kick off on medium sized setups of around 32 - 64 devices. The problem is the use of a connected rados cluster object per image. This only needs to be per cluster and since we only support 1 cluster we only need the one. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
We are changing our implementation for 2.2, so we will not need this fix anymore.