I am using "elasticsearch-cloud-aws" to create backups for my cluster in
S3, but the backups are taking too long to finish, by using cloud watch i
am seeing an average of of 760 MB/minute of network out in the machines, i
used to have similar numbers while the index was balancing itself but by
changing "indices.recovery.max_bytes_per_sec: 1000mb" i got much better
speeds, topping at 5GB/minute.
I would like to know if its there another property that maybe will make the
backup faster or its there a way to make this backups faster.
I am using "elasticsearch-cloud-aws" to create backups for my cluster in
S3, but the backups are taking too long to finish, by using cloud watch i
am seeing an average of of 760 MB/minute of network out in the machines, i
used to have similar numbers while the index was balancing itself but by
changing "indices.recovery.max_bytes_per_sec: 1000mb" i got much better
speeds, topping at 5GB/minute.
I would like to know if its there another property that maybe will make
the backup faster or its there a way to make this backups faster.
Thanks i tested the settings and the speed get better, I also found out ES
1.4.0 has a bug about the snapshot/restore, after upgrading to 1.4.4 my
speed triplicated.
On Monday, March 16, 2015 at 8:44:32 AM UTC-5, Alejandro Calderon wrote:
Hello,
I am using "elasticsearch-cloud-aws" to create backups for my cluster in
S3, but the backups are taking too long to finish, by using cloud watch i
am seeing an average of of 760 MB/minute of network out in the machines, i
used to have similar numbers while the index was balancing itself but by
changing "indices.recovery.max_bytes_per_sec: 1000mb" i got much better
speeds, topping at 5GB/minute.
I would like to know if its there another property that maybe will make
the backup faster or its there a way to make this backups faster.
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.