Description of problem: For an over peta-byte deployment, the product team encouraged us to use the distributed-replicated volume over the dist-dispersed one as they need more rebalance testings with the latter for large-scale deployments. However, we think the dist-dispersed would be best fit for our target use cases, a storage for video streaming, for example. Version-Release number of selected component (if applicable): 3.2 How reproducible: Configure an over PB volume, extend its size by adding nodes and then trigger rebalance. Steps to Reproduce: 1. 2. 3. Actual results: The distributed-replicated volume is recommended. Expected results: We can choose the distributed-dispersed volume as well. Additional info:
Ben, Ambarish, I am aware that one of you are doing the ec perf work with rebalance. Not sure which one it is. Could you take a look at this bug? For now adding needinfo on Ben.
We have done an extensive testing approach to EC with rebalance and other interops too. below is a testplan on what all were covered http://etherpad.corp.redhat.com/3-2-ec-rebalance-plan There were a good number of BZs reported ,which are mentioned in the above doc, which are being tracked seperately Hence moving to verified
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2017-0486.html