Hi,
I have setup a five node cluster (5 shards 1 replica) for evaluation. I
have notice that during a shard relocation or during a cluster rebalance a
node can achieve a maximum speed of 160 Mbp/s.
The issue is that the nodes are communicate through Gigabit Ethernet and
for scp operations i can achieve 560 Mbp/s which is the expected. I haven't
limit the bandwidth from the configuration.
Does anyone has any idea why this might happen?
judging from a quick peek in the source, I think you have to
set indices.recovery.max_bytes_per_sec settings to something more than the
default, which is 20 MB/s (perfectly matching your 160 MBit/s).
--Alex
On Fri, Jun 28, 2013 at 10:36 AM, Antonios Giannopoulos antgiann@gmail.comwrote:
Hi,
I have setup a five node cluster (5 shards 1 replica) for evaluation. I
have notice that during a shard relocation or during a cluster rebalance a
node can achieve a maximum speed of 160 Mbp/s.
The issue is that the nodes are communicate through Gigabit Ethernet and
for scp operations i can achieve 560 Mbp/s which is the expected. I haven't
limit the bandwidth from the configuration.
Does anyone has any idea why this might happen?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.