I want to add ssd1+ssd2 as a disk to elastic-node1 and ssd3+4 to elastic-node2
I also want to add directory from non-ssd /mnt/data for snapshots and warm tier indexes..
Both sound fine. Another option (that I prefer) is to combine your pairs of disks together (using e.g. RAID0 or LVM) and offer a single filesystem to each node, which fits more nicely with the default docker-compose setup that expects a single data path.
Finally my docker file looks like:
docker-compose v2 has cpuset parameter to limit specific CPU, but docker-compose v3 support this only in as part of resource,deploy available only in swarm or docker stack.
The workaround is to use docker-compose --compatibility up -d
the storage i realised to use lvm to create single volume group rather than attaching 2 disk volumes.
the previous option wouldn't work because there are no 2 data directories inside default docker image so we would have to created them before build. Using 1 disk via LVM was easier.
I am not 100% sure if this is good config for production. I do not like --compatibility options and I concern if this would cause some problems in future.
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.