We are observing degraded latencies, we are investigating
Update(s):
Date: 2018-02-22 16:51:20 UTC operation finished, the cluster is stabilized, we will add more hardware in the next days
Date: 2018-02-22 15:49:00 UTC new rebalance in progress
Date: 2018-02-21 15:32:36 UTC we are starting a new rebalance operation
Date: 2018-02-20 18:35:39 UTC the replication method change wasn't applied correctly, we forced a swift-object-replicator restart. Latencies should decrease now
Date: 2018-02-20 16:02:11 UTC replication method switched from rsync to ssync
Date: 2018-02-20 15:41:59 UTC The rebalance is showing its first effects : increased latencies while errors on PUT requests are decreasing. We will switch back replication method to ssync as soon as possible to lower the latencies
Date: 2018-02-20 14:56:17 UTC rebalance in progress
Date: 2018-02-20 13:55:55 UTC We are adding storage nodes to the cluster, a rebalance will follow
Date: 2018-02-19 16:59:36 UTC Due to high dispersion on objects placement we had to switch the replication method from ssync to rsync, which affects performance. Once the dispersion rate reaches an acceptable value we will switch back replication method to ssync